Diagnostic models for impending death in terminally ill cancer patients: A multicenter cohort study.
Ontology highlight
ABSTRACT: Background
Accurately predicting impending death is essential for clinicians to clarify goals of care. We aimed to develop diagnostic models to predict death ≤3 days in cancer patients.Methods
In this multicenter cohort study, we consecutively enrolled advanced cancer patients admitted to 23 inpatient hospices in 2017. Fifteen clinical signs related to impending death were documented daily from the day when the Palliative Performance Scale (PPS) declined to ≤20-14 days later. We conducted recursive partitioning analysis using the entire data set and performed cross-validation to develop the model (prediction of 3-day impending death-decision tree [P3did-DT]). Then, we summed the number of systems (nervous/cardiovascular/respiratory/musculoskeletal), where any sign was present to underpin P3did score (range = 0-4).Results
Data following PPS ≤20 were obtained from 1396 of 1896 inpatients (74%). The mean age was 73 ± 12 years, and 399 (29%) had gastrointestinal tract cancer. The P3did-DT was based on three variables and had four terminal leaves: urine output (u/o) ≤200 ml/day and decreased response to verbal stimuli, u/o ≤200 ml/day and no decreased response to verbal stimuli, u/o >200 ml/day and Richmond Agitation-Sedation Scale (RASS) ≤-2, and u/o >200 ml/day and RASS ≥-1. The 3-day mortality rates were 80.3%, 53.3%, 39.9%, and 20.6%, respectively (accuracy = 68.3%). In addition, 79.6%, 62.9%, 47.2%, 32.8%, and 17.4% of patients with P3did scores of 4, 3, 2, 1, and 0, respectively, died ≤3 days.Conclusion
We successfully developed diagnostic models for death ≤3 days. These may further help clinicians predict impending death and help patients/families prepare for their final days.
SUBMITTER: Mori M
PROVIDER: S-EPMC8607266 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:ObjectivesTo determine the relative influence of different factors on place of death in patients with cancer.Data sourcesFour electronic databases-Medline (1966-2004), PsycINFO (1972-2004), CINAHL (1982-2004), and ASSIA (1987-2004); previous contacts with key experts; hand search of six relevant journals.Review methodsWe generated a conceptual model, against which studies were analysed. Included studies had original data on risk factors for place of death among patients, > 80% of whom had cancer. Strength of evidence was assigned according to the quantity and quality of studies and consistency of findings. Odds ratios for home death were plotted for factors with high strength evidence.Results58 studies were included, with over 1.5 million patients from 13 countries. There was high strength evidence for the effect of 17 factors on place of death, of which six were strongly associated with home death: patients' low functional status (odds ratios range 2.29-11.1), their preferences (2.19-8.38), home care (1.37-5.1) and its intensity (1.06-8.65), living with relatives (1.78-7.85), and extended family support (2.28-5.47). The risk factors covered all groups of the model: related to illness, the individual, and the environment (healthcare input and social support), the latter found to be the most important.ConclusionsThe ne of factors that influence where patients with cancer die is complicated. Future policies and clinical practice should focus on ways of empowering families and public education, as well as intensifying home care, risk assessment, and training practitioners in end of life care.
| S-EPMC1388126 | biostudies-literature
Project description:BackgroundSubstantial interstudy heterogeneity exists in defining acute kidney injury (AKI) and baseline serum creatinine (SCr). This study assessed AKI incidence and its association with pediatric intensive care unit (PICU) mortality under different AKI and baseline SCr definitions to determine the preferable approach for diagnosing pediatric AKI.MethodsIn this multicenter prospective observational cohort study, AKI was defined and staged according to the Kidney Disease: Improving Global Outcome (KDIGO), modified KDIGO, and pediatric reference change value optimized for AKI (pROCK) definitions. The baseline SCr was calculated based on the Schwartz formula or estimated as the upper normative value (NormsMax), admission SCr (AdmSCr) and modified AdmSCr. The impacts of different AKI definitions and baseline SCr estimation methods on AKI incidence, severity distribution and AKI outcome were evaluated.ResultsDifferent AKI definitions and baseline SCr estimates led to differences in AKI incidence, from 6.8 to 25.7%; patients with AKI across all definitions had higher PICU mortality ranged from 19.0 to 35.4%. A higher AKI incidence (25.7%) but lower mortality (19.0%) was observed based on the Schwartz according to the KDIGO definition, which however was overcome by modified KDIGO (AKI incidence: 16.3%, PICU mortality: 26.1%). Furthermore, for the modified KDIGO, the consistencies of AKI stages between different baseline SCr estimation methods were all strong with the concordance rates > 90.0% and weighted kappa values > 0.8, and PICU mortality increased pursuant to staging based on the Schwartz. When the NormsMax was used, the KDIGO and modified KDIGO led to an identical AKI incidence (13.6%), but PICU mortality did not differ among AKI stages. For the pROCK, PICU mortality did not increase pursuant to staging and AKI stage 3 was not associated with mortality after adjustment for confounders.ConclusionsThe AKI incidence and staging vary depending on the definition and baseline SCr estimation method used. The modified KDIGO definition based on the Schwartz method leads AKI to be highly relevant to PICU mortality, suggesting that it may be the preferable approach for diagnosing AKI in critically ill children and provides promise for improving clinicians' ability to diagnose pediatric AKI.
| S-EPMC9264539 | biostudies-literature
Project description:IntroductionUrinary indices have limited effectiveness in separating transient acute kidney injury (AKI) from persistent AKI in ICU patients. Their time-course may vary with the mechanism of AKI. The primary objective of this study was to evaluate the diagnostic value of changes over time of the usual urinary indices in separating transient AKI from persistent AKI.MethodsAn observational prospective multicenter study was performed in six ICUs involving 244 consecutive patients, including 97 without AKI, 54 with transient AKI, and 93 with persistent AKI. Urinary sodium, urea and creatinine were measured at ICU admission (H0) and on 6-hour urine samples during the first 24 ICU hours (H6, H12, H18, and H24). Transient AKI was defined as AKI with a cause for renal hypoperfusion and reversal within 3 days.ResultsSignificant increases from H0 to H24 were noted in fractional excretion of urea (median, 31% (22 to 41%) and 39% (29 to 48%) at H24, P<0.0001), urinary urea/plasma urea ratio (15 (7 to 28) and 20 (9 to 40), P<0.0001), and urinary creatinine/plasma creatinine ratio (50 (24 to 101) and 57 (29 to 104), P=0.01). Fractional excretion of sodium did not change significantly during the first 24 hours in the ICU (P=0.13). Neither urinary index values at ICU admission nor changes in urinary indices between H0 and H24 performed sufficiently well to recommend their use in clinical setting (area under the receiver-operating characteristic curve?0.65).ConclusionAlthough urinary indices at H24 performed slightly better than those at H0 in differentiating transient AKI from persistent AKI, they remain insufficiently reliable to be clinically relevant.
| S-EPMC3733426 | biostudies-literature
Project description:BackgroundThe COVID-19 pandemic caused by SARS-CoV-2 is challenging health care systems globally. The disease disproportionately affects the elderly population, both in terms of disease severity and mortality risk.ObjectiveThe aim of this study was to evaluate machine learning-based prognostication models for critically ill elderly COVID-19 patients, which dynamically incorporated multifaceted clinical information on evolution of the disease.MethodsThis multicenter cohort study (COVIP study) obtained patient data from 151 intensive care units (ICUs) from 26 countries. Different models based on the Sequential Organ Failure Assessment (SOFA) score, logistic regression (LR), random forest (RF), and extreme gradient boosting (XGB) were derived as baseline models that included admission variables only. We subsequently included clinical events and time-to-event as additional variables to derive the final models using the same algorithms and compared their performance with that of the baseline group. Furthermore, we derived baseline and final models on a European patient cohort, which were externally validated on a non-European cohort that included Asian, African, and US patients.ResultsIn total, 1432 elderly (≥70 years old) COVID-19-positive patients admitted to an ICU were included for analysis. Of these, 809 (56.49%) patients survived up to 30 days after admission. The average length of stay was 21.6 (SD 18.2) days. Final models that incorporated clinical events and time-to-event information provided superior performance (area under the receiver operating characteristic curve of 0.81; 95% CI 0.804-0.811), with respect to both the baseline models that used admission variables only and conventional ICU prediction models (SOFA score, P<.001). The average precision increased from 0.65 (95% CI 0.650-0.655) to 0.77 (95% CI 0.759-0.770).ConclusionsIntegrating important clinical events and time-to-event information led to a superior accuracy of 30-day mortality prediction compared with models based on the admission information and conventional ICU prediction models. This study shows that machine-learning models provide additional information and may support complex decision-making in critically ill elderly COVID-19 patients.Trial registrationClinicalTrials.gov NCT04321265; https://clinicaltrials.gov/ct2/show/NCT04321265.
| S-EPMC9015783 | biostudies-literature
Project description:BACKGROUND:Many terminally ill patients in Bosnia-Herzegovina (BiH) fail to receive needed medical attention and social support. In 2016 a primary healthcare centreer (PHCC) in Doboj (BiH) requested the methodological and technical support of a local partner (Fondacija fami) and the Geneva University Hospitals to address the needs of terminally ill patients living at home. In order to design acceptable, affordable and sustainable solutions, we involved patients and their families in exploring needs, barriers and available resources. METHODS:We conducted interviews with 62 purposely selected patients using a semi-structured interview guide designed to elicit patients' experiences, needs and expectations. Both qualitative and quantitative analyses were conducted, using an inductive thematic approach. RESULTS:While patients were aware that their illnesses were incurable, they were poorly informed about medical and social support resources available to them. Family members appeared to be patients' main source of support, and often suffered from exhaustion and financial strain. Patients expressed feelings of helplessness and lack of control over their health. They wanted more support from health professionals for pain and other symptom management, as well as for anxiety and depression. Patients who were bedridden or with reduced mobility expressed strong feelings of loneliness, social exclusion, and stigma from community members and - occasionally - from health workers. CONCLUSIONS:Our findings suggest a wide gap between patients' end-of-life care needs and existing services. In order to address the medical, psychological and social needs of terminally ill patients, a multi-pronged approach is called for, including not only better symptom management through training of health professionals and improved access to medication and equipment, but also a coordinated inter-professional, inter-institutional and multi-stakeholder effort aimed at offering comprehensive medical, psycho-social, educational and spiritual support.
| S-EPMC6245800 | biostudies-other
Project description:BackgroundImmunocompromised critically ill patients constitute a population with the high risk of candidemia. This retrospective study aimed to assess the outcome of immunocompromised critically ill patients with candidemia. Secondary objectives were to describe clinical phenotypes of these patients, Candida ecology, and factors associated with mortality.ResultsOverall, 121 patients were included in this study. Median delay from candidemia to first antifungal therapy was 3 days, in line with the observed delay of blood culture positivity. Candia albicans was the main Candida specie identified (54%), and susceptibility of Candida to fluconazole and echinocandins was of, respectively, 70% and 92%. Hospital mortality was of 60%. After adjustment for confounders, severity as assessed by the need for vasopressors (HR 1.8, CI95% 1.1-3.1), need for mechanical ventilation (HR 2.0, CI95% 1.1-3.8) and allogenic stem cell transplantation (HR 2.5, CI95% 1.1-6.0) were independently associated with poor outcome. Candida specie, susceptibility and treatment strategies were not associated with outcome.ConclusionsCandidemia in immunocompromised critically ill patients is associated with a grim outcome. Despite the high prevalence of Candida non-albicans species, neither C. species nor its susceptibility was associated with outcome. Conversely, severity and preexisting allogeneic stem cell transplantation were independently associated with poor outcome. Despite antifungal prophylaxis and use of preemptive antifungal therapy in neutropenic patients, antifungal therapy was initiated three days after symptoms onset suggesting needs for specific strategies aiming to reduce this delay.
| S-EPMC6546779 | biostudies-literature
Project description:To provide an appropriate method to systematically analyze the hospital discharge of terminally ill patients especially the cooperation between hospital and community nurses and the quality of the discharge handovers. To evaluate the hospital discharge process of terminally ill patients in an academic hospital in the Netherlands using the proposed method.Data were collected from a prospective cohort of all terminally ill patients discharged from the University Medical Center Groningen, the Netherlands, between June and November 2014. The hospital discharges were assessed using 2 questionnaires: an inventory questionnaire, to determine the required care, and an evaluation questionnaire, to evaluate the care actually organized and the discharge handovers. The inventory questionnaire was completed prior to discharge and the evaluation questionnaire between 3 to 7 days after discharge.Around 130 consecutive patients were included. The discharge took place on the desired date in 86% of cases and the average overall discharge grade on a 10-point scale was 7.4 (range: 3-9.5). In 23% of cases discrepancies between required and provided care were identified and medication queries existed in 29%.This study provides a methodology to analyze the hospital discharge procedure of terminally ill patients that can be utlized in any hospital. Structured analysis of the discharge process is valuable and identifies where improvements can be made. Within the study cohort the home care could be arranged at short notice and was considered sufficient. However, in a significant proportion of patients a discrepancy between required and arranged care and queries about medication were identified.
| S-EPMC6257431 | biostudies-literature
Project description:Background For terminally sick neonates and their families, it’s crucial to provide holistic nursing care that incorporates both curative and palliative care as much as feasible. It is well known that the biggest obstacle to delivering palliative care for neonatal children is a lack of training for nurses. Aim: The aim of this research is to investigate the experiences of nurses who provide care for neonates who are terminally ill as well as their educational requirements for neonatal palliative care. Method A cross-sectional descriptive study was conducted among 200 nurses working in a tertiary center providing care for terminally ill neonates in Saudi Arabia. Data was collected from using Neonatal Palliative Care Questionnaire (QNPC) from January 2021 to March 2021. Results Two hundred nurses were surveyed (the response rate was 79%). The mean age of the 158 participants was 35.67 (standard deviation (SD): 7.43), and the majority were female (151; 95.6%). The majority were bachelor’s holders (119; 75.3%), with more than 5 years of experience in providing care for neonates (100; 63.3%). Most of the participants reported not receiving any education about palliative care (115; 72.8%). Nurses reported a moderate level of experience in all areas of neonatal palliative care. The total mean score of palliative care experiences of neonates was 3.42 (SD: 1.35). However, the majority of nurses reported little experience discussing the transition period to palliative care for neonates 2.95 (SD: 1.93), the discussion of code status (DNR) during terminal illness of neonates 3.11 (SD: 1.54) and spiritual support 2.90 (SD: 1.55). Conclusion The assessment of the fundamental skills of neonatal palliative care by nurses was insufficient. To enhance the quality of care, it is crucial to incorporate education on neonatal palliative care into programs for nursing staff development.
| S-EPMC9821778 | biostudies-literature
Project description:ObjectivesTo estimate the incidence, risk factors, and outcomes associated with in-hospital cardiac arrest and cardiopulmonary resuscitation in critically ill adults with coronavirus disease 2019 (covid-19).DesignMulticenter cohort study.SettingIntensive care units at 68 geographically diverse hospitals across the United States.ParticipantsCritically ill adults (age ≥18 years) with laboratory confirmed covid-19.Main outcome measuresIn-hospital cardiac arrest within 14 days of admission to an intensive care unit and in-hospital mortality.ResultsAmong 5019 critically ill patients with covid-19, 14.0% (701/5019) had in-hospital cardiac arrest, 57.1% (400/701) of whom received cardiopulmonary resuscitation. Patients who had in-hospital cardiac arrest were older (mean age 63 (standard deviation 14) v 60 (15) years), had more comorbidities, and were more likely to be admitted to a hospital with a smaller number of intensive care unit beds compared with those who did not have in-hospital cardiac arrest. Patients who received cardiopulmonary resuscitation were younger than those who did not (mean age 61 (standard deviation 14) v 67 (14) years). The most common rhythms at the time of cardiopulmonary resuscitation were pulseless electrical activity (49.8%, 199/400) and asystole (23.8%, 95/400). 48 of the 400 patients (12.0%) who received cardiopulmonary resuscitation survived to hospital discharge, and only 7.0% (28/400) survived to hospital discharge with normal or mildly impaired neurological status. Survival to hospital discharge differed by age, with 21.2% (11/52) of patients younger than 45 years surviving compared with 2.9% (1/34) of those aged 80 or older.ConclusionsCardiac arrest is common in critically ill patients with covid-19 and is associated with poor survival, particularly among older patients.
| S-EPMC7525342 | biostudies-literature
Project description:AimsMidazolam is the drug of choice for palliative sedation and is titrated to achieve the desired level of sedation. Because of large inter-individual variability (IIV), however, the time it takes to achieve adequate sedation varies widely. It would therefore greatly improve clinical care if an individualized dose could be determined beforehand. To find clinically relevant parameters for dose individualization, we performed a pharmacokinetic study on midazolam, 1OH-midazolam (1-OH-M) and 1OH-midazolam-glucuronide (1-OH-MG) in terminally ill patients.MethodsUsing nonlinear mixed effects modelling (NONMEM 7.2), a population pharmacokinetic analysis was conducted with 192 samples from 45 terminally ill patients who received midazolam either orally or subcutaneously. The covariates analysed were patient characteristics, co-medication and blood chemistry levels.ResultsThe data were accurately described by a one compartment model for midazolam, 1-OH-M and 1-OH-MG. The population mean estimates for midazolam, 1-OH-M and 1-OH-MG clearance were 8.4 l h-1 (RSE 9%, IIV 49%), 45.4 l h-1 (RSE 12%, IIV 60.5%) and 5.1 l h-1 (RSE 11%, IIV 49.9%), respectively. 1-OH-MG clearance was correlated with the estimated glomular filtration rate (eGFR) explaining 28.4% of the IIV in 1-OH-MG clearance. In addition, low albumin levels were associated with decreased midazolam clearance, explaining 18.2% of the IIV.ConclusionOur study indicates albumin levels and eGFR as relevant clinical parameters to optimize midazolam dosing in terminally ill patients. The correlation between low albumin levels and decreased midazolam clearance is probably a result of inflammatory response as high CRP levels were correlated in a similar way.
| S-EPMC5510071 | biostudies-literature