Daniel J. Niven

OBJECTIVES: Research examining informed consent in critical care studies is limited. The objective of this study was to describe the patterns and predictors of consent. DESIGN: Research coordinators prospectively documented each consent encounter including consent model (a priori or deferred, also known as consent to continue), who requested consent (research coordinator, site investigator, and ICU physician), who provided or declined consent (patient, substitute decision-maker [SDM], and other), the method to obtain consent (in-person or via telephone), and reasons for consent decline, if applicable. Research coordinators retrospectively reported center characteristics. Using multilevel logistic regression, we evaluated the association of consent provision with characteristics of the consent encounter and participating center. SETTING: International randomized trial of stress ulcer prophylaxis for invasively ventilated patients. INTERVENTION: None. MEASUREMENTS AND MAIN RESULTS: Among 5024 patients or SDMs approached for consent, the consent rate was 85.0% (ranging from 27.3% to 100%). In 2638 patients (52.5%), an a priori approach was used, and of these, 696 patients (26.4%) declined. The remaining 2386 participants (47.5%) were enrolled by deferred consent, of whom 58 participants (2.4%) ultimately declined. Among first consent encounters, 80.2% were with SDMs, 19.6% with patients, and 0.2% with others. Higher consent rates were independently associated with the deferred consent model (odds ratio [OR], 16.56; 95% CI, 11.50–23.84), consent encounters involving patients (OR, 2.67; 95% CI, 1.63–4.37), and consent encounters during the COVID-19 pandemic (OR, 2.01; 95% CI, 1.30–3.11) or postpandemic period (OR, 2.26; 95% CI, 1.34–3.80) compared with prepandemic. Consent rates were not associated with center characteristics including ICU size, ICU research experience, or research coordinator experience. CONCLUSIONS: Deferred consent was used as often as a priori consent in this trial. The informed consent rate was associated with the deferred model, consent from patients, and consent encounters during or after the pandemic.

Objectives: The Coronavirus Disease-2019 (COVID-19) pandemic had indirect impacts on healthcare provided to critically ill intensive care unit (ICU) non-COVID patients and their outcomes. Therefore, we examined the effect of the COVID-19 pandemic on patients with non-COVID critical illness during the pandemic and pre-pandemic epochs. Design: Retrospective, population-based, observational cohort study. Setting: All adult patients admitted to general ICUs in Alberta, Canada, from March 2017–2020 (pre-pandemic period) and March 2020–2023 (pandemic period). Data sources: Data was captured from an integrated critical care clinical information system (eCritical Alberta) and Alberta Health Services (AHS) administrative databases were utilized. Measurements and main results: A total of 80 540 non-COVID patients were admitted to ICUs in Alberta in the period between March 2017 and March 2023, equally distributed between the pandemic (40,196, 50.1%) and pre-pandemic (40,344, 49.9%) periods. For pandemic versus pre-pandemic cohorts, patient mean age was 57.7 ± 16.2 versus 58.5 ± 16.6 years and mean APACHE II score was 19.5 ± 8.8 versus 18.8 ± 8.5, respectively. ICU mortality was higher during the pandemic compared with pre-pandemic (14.2% vs 11.2%, mean difference [MD]: +3.1%, adjusted odds ratio [OR]: 1.21; 95% CI: 1.14-1.28, p < 0.001 ) after logistic regression adjustment. The hospital mortality was significantly higher in the pandemic versus the pre-pandemic period (23.6% vs 15.9%; MD: +7.7%, adjusted OR: 2.06, 95% CI: 1.97-2.16, p < 0.001 ). There was greater ICU (5.95 ± 9.4 vs 5.35 ± 8.5 days, MD: +0.63 days, 95% CI: 0.47 to 0.72 days, p < 0.001), and lower hospital (18.0 ± 28.0 vs 19.9 ± 33.3 days, MD: −1.89 days, 95% CI: −1.45 to −2.32 days p < 0.001) lengths of stay (LOS) during the pandemic compared with pre-pandemic. Conclusion: During the pandemic period, non-COVID patients had worse outcomes, greater adjusted ICU and hospital mortality, with increased ICU lengths of stay, but shorter hospital lengths of stay, when compared with pre-pandemic periods.

Purpose We performed a secondary analysis of a prior systematic review and meta-analysis to examine the prevalence and incidence of ICU delirium and pain. Methods The original search examined MEDLINE, EMBASE, CINAHL, and the Cochrane Central Register of Controlled trials from inception to May 15, 2023 (PROSPERO ID: CRD42022367715). Articles eligible for the current study derived from those previously reviewed at full text review. Included in the current study were randomised or observational studies among critically ill adults, that reported delirium or pain incidence or prevalence. Proportion data was transformed using the Freeman-Tukey double arcsine method prior to pooling using a random effects meta-analysis model. Joanna Briggs’ Institute prevalence checklist was used for risk of bias. Results From 517 full-text articles, 213 original studies, 226 publications (183,285 patients) were included. The pooled delirium prevalence (173 studies) was 35.7% (95%CI 32.4–39.0%). The most common delirium subtype was hypoactive (16.5%, 95%CI 12.1–21.4%). The pooled delirium incidence (41 studies) was 28.8% (95%CI 23.2–34.8%). The pooled pain prevalence (11 studies) was 43.5% (95%CI 28.6–58.9%). Delirium prevalence but not incidence decreased since 2013 (before: 39.9%, 95%CI 34.7–45.3% vs. after: 32.3%, 95%CI 28.2–36.5%, p = 0.02). Pain prevalence decreased significantly since publication of pain, agitation and delirium guidelines in 2013 (before: 64.6%, 95%CI 41.6–84.6% vs. after: 35.8%, 95%CI 20.5–52.6%, p = 0.046). Conclusions Delirium and pain prevalence among adults admitted to ICUs have decreased since publication of pain, agitation and delirium guidelines. However, both remain common, occurring in up to one-half of patients.

Background: Since the beginning of the COVID-19 pandemic, moral distress among healthcare workers in the Intensive Care Unit (ICU) has garnered both media and academic attention. Moral distress has been theorized as occurring when individuals are constrained from doing what they perceive as morally right. This study sought to empirically examine the lived experiences of moral distress among clinical and administrative healthcare professionals in a sample of Canadian ICUs during the COVID-19 pandemic. Methods: Qualitative case study methodology was used as the overarching approach, collecting and comparing data from two distinct cases: one ICU in Ontario and one in Alberta. Data collection involved two primary sources: semi-structured interviews with staff and document review of institutional and government directives to provide contextual data. Data analysis commenced concurrently with data collection, and generated within- and across-case themes, as well as allowed descriptive accounts of moral distress. Results: Thirty-six healthcare workers across two sites were interviewed. Participants described three primary categories of constraints leading to moral distress. These were: 1) The rapidity and opaqueness of policy development, specifically pertaining to 2) the implementation of family visitation and treatment triage decisions, and 3) resource shortages, which reduced patient interactions, shifted professional responsibilities. Each of these constraints yielded circumstances and forced decisions that were perceived as morally wrong because they compromised care quality and outcomes. Conclusions: While sharing similarities with the growing literature on moral distress in the context of the COVID-19 pandemic, this study reveals new insights on how provincial and institutional policy has direct bearing on experiences of moral distress. Policies and circumstances forced ICU staff to choose between actions they considered the most right and the least wrong. Understanding these specific policy-driven constraints highlights the need for healthcare systems and processes that mitigate moral distress and sustain our health workforce.

Purpose We performed a systematic review with meta-analysis examining the relationship between pain or pain medications and delirium occurence, duration, and severity. Methods We searched MEDLINE, EMBASE, CINAHL and the Cochrane Central Register of Controlled Trials from inception to May 15, 2023. We included randomised or observational studies among critically ill adults, that reported data on pain or exposure to analgesics, and reported delirium presence, duration, or severity with no language or region restrictions. PROSPERO ID: CRD42022367715. Two authors independently screened records and extracted data. Risk of bias was evaluated using Risk of Bias 2 or the Risk of Bias In Non-randomized Studies of Interventions. We pooled data using the Hartung-Knapp Sidik-Jonkmann random effects model. PRISMA was followed. Results From 8,054 citations, 90 studies (119,230 patients) published between 2001 and 2023 were included in the systematic review. 41 studies were included in the primary meta-analysis examining prevalent delirium: seven studies evaluated pain; 12 studies evaluated fentanyl; and five studies evaluated morphine. There was a trend to association between pain and delirium occurrence (OR 2.49, 95% CI 0.98–6.30), and a significant association between pain and incident delirium (OR 3.70, 95% CI 1.73–7.93). Fentanyl (OR 2.49, 95% CI 1.45–4.27) and morphine (OR 2.13, 95% CI 1.21–3.75) were associated with delirium occurrence. Risk of bias was critical for many studies. Conclusions We observed an association between pain and incident delirium among critically ill adults. Exposure to morphine or fentanyl (but no other pain medications) was associated with increased risk of delirium occurrence.

IMPORTANCE: Nursing workforce changes, knowledge translation gaps, and environmental/organizational barriers may impact sepsis recognition and management within the ICU. OBJECTIVES: To: 1) evaluate current ICU nursing knowledge of sepsis recognition and management, 2) explore individual and environmental or organizational factors impacting nursing recognition and management of sepsis using the Theoretical Domains Framework (TDF), and 3) describe perceived barriers and facilitators to nursing recognition and management of patients with sepsis. DESIGN, SETTING, AND PARTICIPANTS: This cross-sectional survey was administered to nurses working in four general system ICUs between October 24, 2023, and January 30, 2024. MAIN OUTCOMES AND MEASURES: Quantitative questions (single/multiple choice, true/false, and Likert-based questions eliciting agreement with a statement) were analyzed using descriptive statistics. Open-ended questions exploring barriers and facilitators to sepsis recognition and management were analyzed using qualitative content analysis. RESULTS: A total of 101 completed survey responses were retained. Most nurses agreed early sepsis detection saves lives (n = 98, 97%, TDF domain Beliefs About Consequences) and that nursing care can improve patient outcomes (n = 97, 96%, TDF domain Optimism). Fewer nurses agreed it was easy to identify priority sepsis interventions based on order urgency (n = 53, 53%, TDF domain Memory, Attention, and Decision Processes). Reoccurring barriers and facilitators to sepsis recognition and management were commonly identified across the TDF domains of Knowledge, Skills, Environmental Context and Resources, and Social Influences, including competency deficit (with facilitators including support from colleagues), workload or staffing, and equipment or resource availability. CONCLUSION AND RELEVANCE: ICU nursing sepsis recognition and management is impacted by numerous individual, environmental, and organizational factors. Recommendations include enhanced competency development or support, utilization of structured reinforcement measures (involving the interdisciplinary team and imploring the use of integrative technologies), and addressing equipment/resource-related gaps. Future research and improvement initiatives should use a theory-informed approach to overcome the pervasive, complex challenges impeding timely sepsis recognition and management.

Background: Untreated pain is associated with short-term and long-term consequences, including post-traumatic stress disorder and insomnia. Side effects of some analgesic medications include dysphoria, hallucinations and delirium. Therefore, both untreated pain and analgesic medications may be risk factors for delirium. Delirium is associated with longer length of stay or cognitive impairment. Our systematic review and meta-analysis will examine the relationship between pain or analgesic medications with delirium occurrence, duration and severity among critically ill adults.Methods and analysis: MEDLINE, EMBASE, CINAHL, the Cochrane Central Register of controlled trials and a review of recent conference abstracts will be searched without restriction from inception to 15 May 2023. Study inclusion criteria are: (1) age≥18 years admitted to intensive care; (2) report a measure of pain, analgesic medications and delirium; (3) study design—randomised controlled trial, quasiexperimental designs and observational cohort and case–control studies excluding case reports. Study exclusion criteria are: (1) alcohol withdrawal delirium or delirium tremens; or (2) general anaesthetic emergence delirium; or (3) lab or animal studies. Risk of bias will be assessed with the Risk of Bias V.2 and risk of bias in non-randomised studies tools. There is no language restriction. Occurrence estimates will be transformed using the Freeman-Tukey double arcsine. Point estimates will be pooled using Hartung-Knapp Sidik-Jonkman random effects meta-analysis to estimate a pooled risk ratio. Statistical heterogeneity will be estimated with the I²statistic. Risk of small study effects will be assessed using funnel plots and Egger test. Studies will be analysed for time-varying and unmeasured confounding using E values.Ethics and dissemination: Ethical approval is not required as this is an analysis of published aggregated data. We will share our findings at conferences and in peer-reviewed journals.PROSPERO registration number: The finalised protocol was submitted to the International Prospective Register of Systematic Reviews (PROSPERO ID: CRD42022367715).

Background: Redundancy in laboratory blood tests is common in intensive care units (ICUs), affecting patients’ health and increasing health care expenses. Medical communities have made recommendations to order laboratory tests more judiciously. Wise selection can rely on modern data-driven approaches that have been shown to help identify low-yield laboratory blood tests in ICUs. However, although conditional entropy and conditional probability distribution have shown the potential to measure the uncertainty of yielding an abnormal test, no previous studies have adapted these techniques to include them in machine learning models for predicting abnormal laboratory test results. Objective: This study aimed to address the limitations of previous reports by adapting conditional entropy and conditional probability to extract features for predicting abnormal laboratory blood test results. Methods: We used an ICU data set collected across Alberta, Canada, which included 55,689 ICU admissions from 48,672 patients. We investigated the features of conditional entropy and conditional probability by comparing the performances of 2 machine learning approaches for predicting normal and abnormal results for 18 blood laboratory tests. Approach 1 used patients’ vitals, age, sex, and admission diagnosis as features. Approach 2 used the same features plus the new conditional entropy–based and conditional probability–based features. Both approaches used 4 different machine learning models (fuzzy model, logistic regression, random forest, and gradient boosting trees) and 10 metrics (sensitivity, specificity, accuracy, precision, negative predictive value [NPV], F1 score, area under the curve [AUC], precision-recall AUC, mean G, and index balanced accuracy) to assess the performance of the approaches. Results: Approach 1 achieved an average AUC of 0.86 for all 18 laboratory tests across the 4 models (sensitivity 78%, specificity 84%, precision 82%, NPV 75%, F1 score 79%, and mean G 81%), whereas approach 2 achieved an average AUC of 0.89 (sensitivity 84%, specificity 84%, precision 83%, NPV 81%, F1 score 83%, and mean G 84%). We found that the inclusion of the new features resulted in significant differences for most of the metrics in favor of approach 2. Sensitivity significantly improved for 8 and 15 laboratory tests across the different classifiers (minimum P<.001 and maximum P=.04). Mean G and index balanced accuracy, which are balanced performance metrics, also improved significantly across the classifiers for 6 to 10 and 6 to 11 laboratory tests. The most relevant feature was the pretest probability feature, which is the probability that a test result was normal when a certain number of consecutive prior tests was already normal. Conclusions: The findings suggest that conditional entropy–based features and pretest probability improve the capacity to discriminate between normal and abnormal laboratory test results. Detecting the next laboratory test result is an intermediate step toward developing guidelines for reducing overtesting in the ICU.