Danish Intensive Care Units Research Articles

Prevalence and etiology of ventilator-associated pneumonia during the COVID-19 pandemic in Denmark: Wave-dependent lessons learned from a mixed-ICU.

Ventilator-associated pneumonia (VAP) may be a particular concern in patients with severe coronavirus disease 2019 (COVID-19). We aimed to determine the prevalence and etiology of VAP in critically ill COVID-19 patients in a Danish intensive care unit (ICU) during the first three waves of the COVID-19 pandemic and to study associations between dexamethasone (DXM) use and development of VAP. In an observational single-center study patients were retrospectively screened for VAP including causative pathogens, use of DXM and commonly used antibiotics. Diagnosis of VAP required invasive mechanical ventilation (IMV) >48 h with presence of a new bacterial agent and clinical signs of infection. For analysis, common descriptive statistics were applied. Cox proportional hazards models were used to analyze the association between DXM use and VAP. VAP was detected in 53/119 (44.5%) mechanically ventilated patients across all three COVID-19 waves. Median length of IMV for VAP patients was 24 [15-41] days, and 3 out of 4 were males. VAP was most prevalent (47.0%) during the second wave. Common pathogens included Klebsiella pneumoniae (24.5%), Enterobacter aerogenes (17.0%) and Pseudomonas aeruginosa (13.2%), Staphylococcus aureus (13.2%), and Escherichia coli (13.2%). A change from Gram-negative bacteria only to a combination of Gram-positive and Gram-negative bacteria was observed in the second wave compared to first. Use of DXM was not associated with VAP (adjusted hazard ratio 1.63 95% CI: 0.84-3.17). The prevalence of VAP was high across all three COVID-19 waves and showed a different distribution of pathogens between the first and second wave. Use of DXM was not associated with VAP development. Further and larger studies are needed to understand the risk factors associated with VAP in patients with COVID-19.

A window of opportunity for ICU end-of-life care-A retrospective multicenter cohort study.

The "window of opportunity" for intensive care staff to deliver end-of-life (EOL) care lies in the timeframe from "documenting the diagnosis of dying" to death. Diagnosing the dying can be a challenging task in the ICU. We aimed to describe the trajectories for dying patients in Danish intensive care units (ICUs) and to examine whether physicians document that patients are dying in time to perform EOL care and, if so, when a window of opportunity for EOL care exists. From the Danish Intensive Care Database, we identified patients ≥18 years old admitted to Danish ICUs between January and December 2020 with an ICU stay of >96 h (four days) and who died during the ICU stay or within 7 days after ICU discharge. A chart review was performed on 250 consecutive patients admitted from January 1, 2020, to ICUs in the Central Denmark Region. In most charts (223 [89%]), it was documented that the patient was dying. Of those patients who received mechanical ventilation, 171 (68%) died after abrupt discontinuation of mechanical ventilation, and 63 (25%) died after gradual withdrawal. Patients whose mechanical ventilation was discontinued abruptly died after a median of 1 h (interquartile range [IQR]: 0-15) and 5 h (IQR: 2-15) after a diagnosis of dying was recorded. In contrast, patients with a gradual withdrawal died after a median of 108 h (IQR: 71-189) and 22 h (IQR: 5-67) after a diagnosis of dying was recorded. EOL care hinges on the ability to diagnose the dying. This study shows that there is a window of opportunity for EOL care, particularly for patients who are weaned from mechanical ventilation. This highlights the importance of intensifying efforts to address EOL care requirements for ICU patients and those discharged from ICUs who are not eligible for readmission.

Veno-venous extracorporeal membrane oxygenation for severe COVID-19 associated acute respiratory distress syndrome: A retrospective, nationwide, Danish cohort study.

Severe acute respiratory syndrome (ARDS) may require veno-venous extracorporeal membrane oxygenation (V-V ECMO). The aim of this study was to provide data on patient selection and outcome in a nationwide cohort study of patients with COVID-19 associated ARDS supported with V-V ECMO. We identified all patients with COVID-19, who were supported with V-V ECMO in Denmark from March 10, 2020, to December 31, 2021, and retrieved data on patients who were referred to- and accepted for ECMO, demographics, outcome data, and complications. Risk factors for mortality were analysed using multivariate Cox regression analysis. During the study period, 1836 patients were admitted to Danish intensive care units (ICUs). In the same period, there were 197 enquiries for ECMO of whom 118 patients were considered eligible. Overall, 71 patients were cannulated for ECMO; three patients were cannulated for veno-arterial extracorporeal membrane oxygenation (V-A ECMO) due to right sided heart failure and 68 patients were cannulated for V-V ECMO. Two patients accepted for V-V ECMO died during cannulation. The median age was 55 years (IQR 45-60) and 66% were males. The median duration of ECMO support was 13 days (IQR 7-21), mechanical ventilation median 26 days (IQR 14-42), ICU stay median 34 days (IQR 17-46), and length of hospital stay median 41 days (IQR 25-56). Ninety-day mortality was 43%. Age of 60 years or more was associated with an increased risk of mortality. Pre-existing hypertension was associated with a decreased risk of mortality. A nationwide, Danish cohort study of 68 COVID-19 patients supported with V-V ECMO, showed a 90-day survival of 43%, which is in accordance with reports from comparable cohorts. Age of 60 years or more was associated with an increased risk of mortality. Pre-existing hypertension was associated with a decreased risk of mortality.

Postextubation dysphagia management in Danish intensive care units: A national survey.

Postextubation dysphagia (PED) is a common complication to endotracheal intubation in critically ill patients and may lead to pneumonia, prolonged ventilation, longer hospital stays, and increased mortality. Recognizing dysphagia is paramount to preventing adverse events. The aim of this study was to describe PED management by investigating practice in Danish intensive care units (ICUs) focusing on current practice in 2023 (screening, prevention, and treatment), perceived best practice (barriers and facilitators), and when possible, to compare practice in 2017 and 2023. Self-reported, cross-sectional survey of dysphagia practice in Danish ICUs administered from April to May 2023. In addition, data were compared with the 2017 Dysphagia in Intensive Care Evaluation study, when possible. Only half of Danish ICUs reported to have a PED protocol, and less than half routinely screen patients for dysphagia after extubation. Most common screening methods were the oral mechanism examination, water test, and Facio-oral tract therapy. Nurses and physicians often relied on an overall physical assessment of the patient. Best treatment methods were uniformly agreed to be patient positioning, modification of food and fluids, use of ergonomic utensils, and compensatory maneuvers. Key barriers to dysphagia management were lack of specialized staff, under-recognition of dysphagia as a health issue, and lack of standardized protocols. Awareness of PED is increasing and identification, prevention, and treatment is slowly improving, but systematic implementation of protocols for dysphagia screening and treatment could enhance dysphagia management in Danish ICUs.

Making It Count – Tracing Signs of Consciousness and Potentiality in Severe Brain Injury in Denmark

ABSTRACT Healthcare professionals use various technologies to evaluate and support patients who have suffered severe brain injuries. They integrate monitoring and sensory assessments into their clinical practice, and these assessments can have an impact on treatment decisions and prognostication. Responses from patients during different interactions are interpreted as “signs of consciousness” when considered contextually relevant. This study is based on anthropological fieldwork conducted in specialized Danish intensive care units, where we explore how signs of consciousness are made to count through practices of enactment. We ethnographically trace how the clinical concept of potential influences the interpretation of signs of consciousness as a complex biosocial practice based on the biomedical assumption that consciousness is a vital indicator of what makes a life. The article provides insights into the potential for recovery as an emergent biosocial practice and contributes to a broader discussion within medical anthropology of the moral landscapes of clinical and experimental borderlands.

Delirium and delirium severity screening in the intensive care—correspondence of screenings tools

BackgroundDelirium severity scores are gaining acceptance for measuring delirium in the intensive care unit (ICU). ObjectiveThe aim of this study was to evaluate the concordance between the Confusion Assessment Method for the intensive care unit (CAM-ICU)-7 and the Intensive Care Delirium Screening Checklist (ICDSC) as delirium severity measurement tools. MethodsThis was a prospective, comparative, observational multicentre study. This study was conducted in 18 Danish ICUs. Delirium was assessed in adult critically ill patients admitted to an ICU with a Richmond Agitation and Sedation Score (RASS) of −2 or above. ICU nurses assessed delirium with randomised paired delirium screening instruments, using the CAM-ICU, the ICDSC, and the CAM-ICU-7. The correlation between the CAM-ICU-7 and the ICDSC severity scores was evaluated for all predefined patient subgroups. ResultsA total of 1126 paired screenings were conducted by 127 ICU nurses in 850 patients. The patients' median age was 70 years (interquartile range: 61–77), 40% (339/850) were female, and 54% (457/850) had at least one positive delirium score. Delirium severity ranges (CAM-ICU-7: 0–7; and ICDSC: 0–8) were positively correlated (Pearson's correlation coefficient, r = 0.83; p < 0.0001). The overall agreement between the CAM-ICU-7 and the ICDSC for delirium measurement (CAM-ICU-7: >2, and ICDSC: >3) was substantial (kappa = 0.74), but the agreement decreased to fair (kappa = 0.38) if a patient had a RASS less than 0. ConclusionsThe agreement between the CAM-ICU-7 and the ICDSC for delirium severity measurement was substantial but might be affected by the patient's sedation and agitation level at the time of assessment. Implications for practiceBoth CAM-ICU-7 and ICDSC can be implemented for delirium severity measurement. Attention is warranted in both scores if a patient has a RASS of −2.

Patient and family engagement in Danish intensive care units: A national survey.

Patient and family engagement in the intensive care unit increases the quality of care and patient safety. The aim of our study was to describe current practice and experiences of contemporary patient and family engagement in the intensive care unit at the individual level, the organizational level, and in the research process according to critical care nurses. We conducted a national qualitative survey of intensive care units in Denmark from 5th May-5th June 2021. Questionnaires were piloted and sent to intensive care nurse specialists and research nurses at 41 intensive care units, allowing one respondent per unit. All respondents were provided with written information about the study by email, and by activating the survey link, they accepted participation. Thirty-two nurses responded to the invitation, 24 completed and 8 partially completed the survey, yielding a response rate of 78%. At the individual level, 27 respondents stated that they involved patients and 25 said they involved family in daily treatment and care. At the organizational level, 28 intensive care units had an overall strategy or guideline for patient and family engagement, and 4 units had established a PFE panel. And, finally, 11 units engaged patients and families in the research process. Our survey suggested that patient and family engagement was implemented to some degree at the individual level, organizational level, and in the research process, but only 4 units had established a PFE panel at the organizational level, which is key to engagement. Patient engagement increases when patients are more awake, and family engagement increases when patients are unable to participate. Engagement increases when patient and family engagement panels are implemented.

Abnormal serum levels of magnesium, phosphate, and zinc in ICU patients-Characteristics, management, and outcomes: The WhyTrace cohort study.

Abnormal serum levels of magnesium, phosphate, and zinc appear common in intensive care unit (ICU) patients, but the epidemiology, management, and associations with outcomes are less well described. We described these factors and estimated associations with outcomes in a large dataset of Danish ICU patients. We included adults who were acutely admitted to 10 general ICUs in Denmark between October 2011 and January 2018. From the dataset, we obtained characteristics of patients who had serum levels measured of magnesium, phosphate, or zinc, including data on supplementation. We used joint models with death as a competing outcome to estimate the associations between abnormal serum levels and time to successful extubation and, for magnesium, also incident tachyarrhythmia. We included 16,517 of 36,514 patients in the dataset. The cumulative probability of hypomagnesemia within 28 days was 64% (95% confidence interval [CI] 62-66); of hypophosphatemia 74% (95% CI 72-75); and of hypozincemia 98% (95% CI 98-98). Supplementation of magnesium was used in 3554 out of 13,506 (26%) patients, phosphate in 2115 out of 14,148 (15%) patients, and zinc in 4465 out of 9869 (45%) patients. During ICU stay, 38% experienced hypermagnesemia, 58% hyperphosphatemia, and 1% hyperzincemia. Low serum levels of magnesium, phosphate, and zinc were associated with shorter time to successful extubation, and high serum magnesium and phosphate and low serum zinc with the competing risk of increased mortality, but too few serum measurements made the results inconclusive. In this multicenter cohort study of acutely admitted ICU patients, most experienced low serum levels of magnesium, phosphate, or zinc during ICU stay, many received supplementation, and experiencing both low and high serum levels during ICU stay was not uncommon. Associations between serum levels and clinical outcomes appeared inconclusive because the data proved unfit for these analyses.

Effects of 12 mg vs. 6 mg dexamethasone on thromboembolism and bleeding in patients with critical COVID-19 - a post hoc analysis of the randomized, blinded COVID STEROID 2 trial

BackgroundThromboembolism is more common in patients with critical COVID-19 than in other critically ill patients, and inflammation has been proposed as a possible mechanism. The aim of this study was to investigate if 12 mg vs. 6 mg dexamethasone daily reduced the composite outcome of death or thromboembolism in patients with critical COVID-19.MethodsUsing additional data on thromboembolism and bleeding we did a post hoc analysis of Swedish and Danish intensive care unit patients enrolled in the blinded randomized COVID STEROID 2 trial comparing 12 mg vs. 6 mg dexamethasone daily for up to 10 days. The primary outcome was a composite outcome of death or thromboembolism during intensive care. Secondary outcomes were thromboembolism, major bleeding, and any bleeding during intensive care.ResultsWe included 357 patients. Whilst in intensive care, 53 patients (29%) in the 12 mg group and 53 patients (30%) in the 6 mg group met the primary outcome with an unadjusted absolute risk difference of − 0.5% (95% CI − 10 to 9.5%, p = 1.00) and an adjusted OR of 0.93 (CI 95% 0.58 to 1.49, p = 0.77). We found no firm evidence of differences in any of the secondary outcomes.ConclusionsAmong patients with critical COVID-19, 12 mg vs. 6 mg dexamethasone daily did not result in a statistically significant difference in the composite outcome of death or thromboembolism. However, uncertainty remains due to the limited number of patients.

Nurses' experiences of the phenomenon 'isolation communication'.

Communication with patients and relatives can be a challenge in the intensive care unit (ICU) setting because of the acute and critical situation. However, when caring for patients with infectious diseases like COVID-19, nurses' communication is challenged further due to the required use of personal protective equipment (PPE) and mandatory isolation regimes. To explore nurses' experiences of communicating while wearing PPE during COVID-19 isolation precautions in the ICU. A qualitative study within the phenomenological-hermeneutic tradition. Data were collected via 12 interviews with nurses working in a Danish ICU from September to November 2020. Data were analysed using a Ricoeur-inspired text interpretation method. Three themes emerged during the analysis: (1) communication was limited and distanced and nurses had to compromise; (2) the nurses' senses were reduced, and verbal and nonverbal communication practises changed; and (3) patients' and relatives' communicative vulnerability were exposed in an extraordinary situation. The analysis revealed a phenomenon that can be described as 'isolation communication'. The isolation precautions and use of PPE had a profound impact on the nurses' caring and communicative practices, which were limited in this situation. The nurses found themselves physically, emotionally and socially distanced from the patients, relatives and their colleagues. However, to prevent the spread of the virus, isolation communication is something that the nurses have to endure. As our findings show that the nurses' communicative practises had to change during isolation communication, it is important for clinical practise and education to focus on implementing communication methods that optimise message transmission between ICU clinicians, patients and relatives in conditions requiring PPE and isolation. We should also focus on how to optimise interdisciplinary health communication in this situation.

Changes over time in characteristics, resource use and outcomes among ICU patients with COVID-19-A nationwide, observational study in Denmark.

BackgroundCharacteristics and care of intensive care unit (ICU) patients with COVID‐19 may have changed during the pandemic, but longitudinal data assessing this are limited. We compared patients with COVID‐19 admitted to Danish ICUs in the first wave with those admitted later.MethodsAmong all Danish ICU patients with COVID‐19, we compared demographics, chronic comorbidities, use of organ support, length of stay and vital status of those admitted 10 March to 19 May 2020 (first wave) versus 20 May 2020 to 30 June 2021. We analysed risk factors for death by adjusted logistic regression analysis.ResultsAmong all hospitalised patients with COVID‐19, a lower proportion was admitted to ICU after the first wave (13% vs. 8%). Among all 1374 ICU patients with COVID‐19, 326 were admitted during the first wave. There were no major differences in patient's characteristics or mortality between the two periods, but use of invasive mechanical ventilation (81% vs. 58% of patients), renal replacement therapy (26% vs. 13%) and ECMO (8% vs. 3%) and median length of stay in ICU (13 vs. 10 days) and in hospital (20 vs. 17 days) were all significantly lower after the first wave. Risk factors for death were higher age, larger burden of comorbidities (heart failure, pulmonary disease and kidney disease) and active cancer, but not admission during or after the first wave.ConclusionsAfter the first wave of COVID‐19 in Denmark, a lower proportion of hospitalised patients with COVID‐19 were admitted to ICU. Among ICU patients, use of organ support was lower and length of stay was reduced, but mortality rates remained at a relatively high level.

The use of melatonin in Danish intensive care departments-A nationwide observational study.

Melatonin is widely employed as a hypnotic in various patient groups. In intensive care patients, melatonin seems to be increasingly used due to potential clinical effects and a favourable safety profile. We aimed to investigate the extend of usage and clinical practice of melatonin therapy in intensive care departments in Denmark. Data from regional hospital pharmacies and the Danish Intensive Care Database were used to estimate defined daily dose and defined daily dose per 1000 ICU admission days. Also, related expenses in the period 2015-2019. Finally, a questionnaire describing the clinical practice of melatonin therapy was provided to all Danish intensive care departments. The usage of melatonin in intensive care departments in Denmark increased from 21,300 DDD (200.0 DDD per 1000 ICU admission days) in 2015 to 52,170 DDD (560.7 DDD per 1000 ICU admission days) in 2019. A total of 32 ICU departments participated in the study (97% of all Danish ICU departments). All included ICU departments employed melatonin as a hypnotic. Nineteen percent of included departments administered melatonin to all admitted patients, whereas 25% of departments rarely administered melatonin. Magistral melatonin 3-mg tablets was the most employed drug dose/formulation. Increased doses of melatonin were administered in selected patients. Melatonin was considered safe by prescribing clinicians. Melatonin is widely and increasingly used in Danish intensive care departments. The more than doubled usage of melatonin in the study period advocates for further studies employing validated outcomes of sleep and other patient-relevant outcomes. This study documents that melatonin is frequently used as a hypnotic in Danish intensive care units during recent years despite a shortage of reliable evidence to support a recommendation to treat with melatonin in this context. These results support a need for conducting clinical trials to determine whether or not there is a beneficial effect of melatonin treatment in critically ill patients.

Public versus patient health preferences: protocol for a study to elicit EQ-5D-5L health state valuations for patients who have survived a stay in intensive care

IntroductionThe value set used when calculating quality-adjusted life-years (QALYs) is most often based on stated preference data elicited from a representative sample of the general population. However, having a severe...

Elements of chaplaincy in Danish intensive care units: key-informant interviews with hospital chaplains

Critically ill patients often require spiritual counseling. In Denmark, most hospital chaplains are provided by the State Church. More knowledge is needed regarding health care chaplaincy in this semi-agnostic non-secular welfare society. Our study aimed to explore how hospital chaplains describe their role and elements of chaplaincy in relation to patient, family and healthcare professionals in Danish intensive care units. Using a multicenter qualitative design, we performed semi-structured key-informant interviews with ten hospital chaplains. Content analysis exposed elements of chaplaincy based on spirituality, rituality, neutrality, confidentiality, advocacy, and equality. Most elements are similar in other countries; the main differences are related to educational requirements, institutional integration, research activity, and documentation practice. Chaplains fill an ambiguous role, negotiating values of church and hospital. Modern chaplaincy is propelled toward professionalism, quantification, certification and evaluation. We discuss the contemporary role of hospital chaplains and the consequences of categorizing them as part of the ICU team.

Translation and adaption of the Yale Swallow Protocol for a Danish intensive care setting.

Dysphagia related to intubation affects almost one in five patients in the intensive care unit; it may vary due to fluctuations in cerebral status and fatigue and lead to pneumonia, re-intubation and increased mortality. Taking care of patients' nutritional needs is an integral part of nursing and nurses must therefore be able to assess risk of dysphagia at all hours. However, Danish validated methods for use in the critically ill patient population are lacking. To translate and adapt the Yale Swallow Protocol method for implementation in a Danish intensive care unit setting. Translation and adaption followed WHO's recommendations for translation and adaption of instruments. Face validity of the Danish version was evaluated in four group interviews. These were digitally recorded, transcribed and analysed using deductive content analysis. All interviewees gave informed consent. The translated version of the Yale Swallow Protocol was easy to understand, although nurses needed to have the method explained and needed training in using the method. In addition, the Yale Swallow Protocol must be adapted to local practice in terms of identifying patients at risk of dysphagia and deciding where to document results of the screening. The translated version of the Yale Swallow Protocol proved conceptually and culturally sound and acceptable to intensive care unit nurses. Implementation should be followed by theoretical and practical training to build a nursing vocabulary relating to dysphagia and strengthen nurses' sensitivity to clinical manifestations consistent with dysphagia.

Characteristics, interventions, and longer term outcomes of COVID-19 ICU patients in Denmark-A nationwide, observational study.

Most data on intensive care unit (ICU) patients with COVID-19 originate in selected populations from stressed healthcare systems with shorter term follow-up. We present characteristics, interventions and longer term outcomes of the entire, unselected cohort of all ICU patients with COVID-19 in Denmark where the ICU capacity was not exceeded. We identified all patients with SARS-CoV-2 admitted to any Danish ICU from 10 March to 19 May 2020 and registered demographics, chronic comorbidities, use of organ support, length of stay, and vital status from patient files. Risk factors for death were analyzed using adjusted Cox regression analysis. There were 323 ICU patients with confirmed COVID-19. Median age was 68years, 74% were men, 50% had hypertension, 21% diabetes, and 20% chronic pulmonary disease; 29% had no chronic comorbidity. Invasive mechanical ventilation was used in 82%, vasopressors in 83%, renal replacement therapy in 26%, and extra corporeal membrane oxygenation in 8%. ICU stay was median 13days (IQR 6-22) and hospital stay 19days (11-30). Median follow-up was 79days. At end of follow-up, 118 had died (37%), 15 (4%) were still in hospital hereof 4 in ICU as of 16 June 2020. Risk factors for mortality included male gender, age, chronic pulmonary disease, active cancer, and number of co-morbidities. In this nationwide, population-based cohort of ICU patients with COVID-19, longer term survival was high despite high age and substantial use of organ support. Male gender, age, and chronic co-morbidities, in particular chronic pulmonary disease, were associated with increased risk of death.

Risk factors for long-term cognitive impairment in ICU survivors: A multicenter, prospective cohort study.

To describe the incidence of and risk factors for impaired cognitive function in intensive care unit (ICU) survivors. We hypothesized that age, severity of illness, and days in coma, delirium, mechanical ventilation in the ICU would be associated with impaired cognitive function. We included all adults, alive 6months after acute admission to one of the 24 Danish ICUs participating in the AID-ICU cohort study. Trained professionals assessed cognitive function in patients' homes or in outpatient clinics using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) 6months after ICU admission. Potential risk factors for cognitive impairment were analyzed with linear regression models. In total, 237 ICU patients were alive 6months after ICU admission and did not meet the exclusion criteria. A total of 106 patients completed the cognitive assessment. The median RBANS global cognitive score was 76 (interquartile range, 62-91), and 52% had a global cognitive score 1.5 SD below the normative mean and 36% displayed a global cognitive score 2 SD below the normative mean, similar to that of Alzheimer's disease. Higher age was associated with poorer RBANS global cognitive score (estimate -0.35 [95% confidence interval -0.63 to -0.07] per year). In this multicenter study of adult ICU survivors, cognitive impairment was frequent and severe in those assessed at 6months. Higher age was a risk factor for cognitive impairment, but events related to the ICU stay were not associated with poorer cognitive performance at 6months.

Use of red blood cells in Danish intensive care units: A population-based register study.

Red blood cell (RBC) transfusion is common in the intensive care unit (ICU). Recent trials have shown that a restrictive transfusion strategy is safe in most patients, and recent guidelines recommend such a strategy in most ICU patients. It is unknown if this has translated into a change in clinical practice. We conducted a population-based register study of RBC transfusions in ICUs in the Danish Capital Region between 1st ofJanuary 2011 and 31stof December 2016 by linking data from the regional blood bank and the Danish Intensive Care Database. We used crude data and run- and control-charts to analyse changes in the number of RBC transfusions. We included 27835 ICU admissions of which 6936 received 40889 RBC units. The crude use was 36.2 RBC units per one-hundred patient bed-days in 2011 vs 29.8 in 2016. The run-chart analysis did not confirm a change in the total use of RBC units in all ICUs combined, and we observed no change in the proportion of transfused patients or in the use of RBCs among transfused patients. Sensitivity analyses showed decreased use of RBC units in two general ICUs, and a reduced use of RBC units among medical ICU patients. In this population-based register study, we did not with certainty observe changes over time in the use of RBC transfusions in all patients in all ICUs in the Danish Capital Region. A reduction in RBC use may have occurred in some general ICUs and in medical ICU patients.

Influenza vaccination and 1-year risk of myocardial infarction, stroke, heart failure, pneumonia, and mortality among intensive care unit survivors aged 65years or older: a nationwide population-based cohort study.

We examined whether influenza vaccination affects 1-year risk of myocardial infarction, stroke, heart failure, pneumonia, and death among intensive care unit (ICU) survivors aged ≥ 65years. Danish Intensive Care Database data on all elderly ( ≥ 65 years) patients hospitalized in Danish ICUs in the period 2005-2015, and subsequently discharged, were linked with data from other medical registries, including data on uptake of the seasonal influenza vaccine. We computed these patients' 1-year risk of hospitalization for myocardial infarction, stroke, heart failure, or pneumonia, and their 1-year risk of all-cause mortality. Hazard ratios (HRs) with 95% confidence intervals (CIs) were computed using Cox proportional hazards regression, with adjustment and propensity score matching applied to handle confounding. The study included 89,818 ICU survivors. The influenza vaccinated patients (n = 34,871, 39%) were older, had more chronic diseases, and used more prescription medications than the unvaccinated patients. Adjusted 1-year mortality was decreased among the vaccinated versus the unvaccinated patients (19.3% versus 18.8%; adjusted HR, 0.92; 95% CI 0.89-0.95). Influenza vaccination was also associated with a decreased risk of stroke (adjusted HR, 0.84; 95% CI 0.78-0.92), but only a small, non-significantly decreased risk of myocardial infarction (adjusted HR, 0.93; 95% CI 0.83-1.03). There was no association between vaccination and subsequent hospitalization for heart failure or pneumonia. Propensity score matched analyses confirmed these findings. Compared with the unvaccinated ICU survivors, the influenza vaccinated ICU survivors had a lower 1-year risk of stroke and a lower 1-year risk of death, whereas no substantial association was observed for the risk of hospitalization for myocardial infarction, heart failure, or pneumonia. Our findings support influenza vaccination of individuals aged ≥ 65years.

Survival prediction in intensive-care units based on aggregation of long-term disease history and acute physiology: a retrospective study of the Danish National Patient Registry and electronic patient records

Intensive-care units (ICUs) treat the most critically ill patients, which is complicated by the heterogeneity of the diseases that they encounter. Severity scores based mainly on acute physiology measures collected at ICU admission are used to predict mortality, but are non-specific, and predictions for individual patients can be inaccurate. We investigated whether inclusion of long-term disease history before ICU admission improves mortality predictions. Registry data for long-term disease histories for more than 230 000 Danish ICU patients were used in a neural network to develop an ICU mortality prediction model. Long-term disease histories and acute physiology measures were aggregated to predict mortality risk for patients for whom both registry and ICU electronic patient record data were available. We compared mortality predictions with admission scores on the Simplified Acute Physiology Score (SAPS) II, the Acute Physiologic Assessment and Chronic Health Evaluation (APACHE) II, and the best available multimorbidity score, the Multimorbidity Index. An external validation set from an additional hospital was acquired after model construction to confirm the validity of our model. During initial model development data were split into a training set (85%) and an independent test set (15%), and a five-fold cross-validation was done during training to avoid overfitting. Neural networks were trained for datasets with disease history of 1 month, 3 months, 6 months, 1 year, 2·5 years, 5 years, 7·5 years, 10 years, and 23 years before ICU admission. Mortality predictions with a model based solely on disease history outperformed the Multimorbidity Index (Matthews correlation coefficient 0·265 vs 0·065), and performed similarly to SAPS II and APACHE II (Matthews correlation coefficient with disease history, age, and sex 0·326 vs 0·347 and 0·300 for SAPS II and APACHE II, respectively). Diagnoses up to 10 years before ICU admission affected current mortality prediction. Aggregation of previous disease history and acute physiology measures in a neural network yielded the most precise predictions of in-hospital mortality (Matthews correlation coefficient 0·391 for in-hospital mortality compared with 0·347 with SAPS II and 0·300 with APACHE II). These results for the aggregated model were validated in an external independent dataset of 1528 patients (Matthews correlation coefficient for prediction of in-hospital mortality 0·341). Longitudinal disease-spectrum-wide data available before ICU admission are useful for mortality prediction. Disease history can be used to differentiate mortality risk between patients with similar vital signs with more precision than SAPS II and APACHE II scores. Machine learning models can be deconvoluted to generate novel understandings of how ICU patient features from long-term and short-term events interact with each other. Explainable machine learning models are key in clinical settings, and our results emphasise how to progress towards the transformation of advanced models into actionable, transparent, and trustworthy clinical tools. Novo Nordisk Foundation and Innovation Fund Denmark.