The Very Elderly Admitted to ICU: A Quality Finish?
The Very Elderly Admitted to ICU: A Quality Finish?
We conducted a multicenter, prospective, observational cohort study with a nested cohort from September 2009 to February 2013 in 24 Canadian ICUs (for participating centers, see Appendix 1, Supplemental Digital Content 1, http://links.lww.com/CCM/B299). We report here an analysis of the ICU treatments and hospital outcomes in all patients (hospital cohort). Details of long-term outcomes of the nested cohort have been reported elsewhere (Heyland et al, unpublished observations, 2015). Institutional research ethics board approval was obtained from each center. Written informed consent was obtained from each patient's next of kin before enrollment.
In participating centers, all patients aged 80 or more were eligible for the hospital outcomes cohort; inclusion was based on research coordinator availability and resources to support the data collection. We obtained a waiver of informed consent to collect data that was readily available from the hospital charts. However, patients in the hospital were eligible for enrollment into a nested cohort study for which consent was required. An eligible family member was required for participation in the follow-up study. Patients were excluded from the nested cohort if they were imminently dying, were not expected to stay in ICU for more than 24 hours, had no permanent address, or were not Canadian residents (due to anticipated challenges in posthospital follow-up). An eligible family member consisted of a family member including partner, significant other, and/or close friend who: 1) provided the most care to the patient before hospital admission and was not paid to do so; 2) visited the patient at least once during the index ICU admission; and 3) was 18 years old or older. We also excluded patients from the nested cohort who did not have a French- or English-speaking family member.
Baseline demographic and clinical data collected from the medical chart for all patients included: age, sex, ICU admission diagnosis, admission type (medical vs surgical), Acute Physiology and Chronic Health Evaluation II score, functional comorbidity index, and Charlson comorbidity index. We assessed sequential organ failure assessment scores at admission and daily while in ICU.
For patients who were enrolled in the nested follow-up, research coordinators met with the family member to collect additional data, including a measure of baseline frailty using the Clinical Frailty Scale. This score ranges from 1 (very fit) to 7 (very frail) and has been previously validated in the ICU. In addition, we documented whether the patient had a living will or advance directive about preferred medical care if he/she was seriously ill. We did not record the contents of the advance directive; just whether the patient had completed one. Finally, we asked the patient's family member which of the following three options they preferred for their loved one: 1) "Life Support. This option can include the use of a breathing machine as well as drugs and procedures to maintain body functions. Efforts are made to keep the patient as comfortable as possible." 2) "Comfort Care without Life Support. Comfort care aims to relieve suffering and preserve the dignity of the patient, without prolonging the dying process. There are usually fewer tests, fewer tubes, and no life support machines or monitors connected to the patient." 3) "I am unsure."
Use of life-sustaining treatment (mechanical ventilation, vasoactive drugs, and renal replacement therapy) and ICU and hospital length of stay were abstracted from the medical record in all patients. Medical orders for administration, withholding, or withdrawing of life-sustaining treatment before ICU admission and during ICU stay were also recorded. Withholding life-sustaining treatment was defined as no current receipt of any such treatment followed by a written order not to start or restart it. Withdrawing life-sustaining treatment was defined as current receipt of any such treatment followed by a written order to discontinue it in anticipation of death.
Our primary outcome was a prolonged dying experience defined arbitrarily as staying in ICU for 7 days or longer and dying during the hospitalization. Baseline patient characteristics, treatments, and outcomes were compared between patients who 1) died within 7 days of ICU admission, 2) died at any point in hospital after staying in ICU for at least 7 days, and 3) survived to hospital discharge. In addition to the overall analyses, in the nested cohort, we conducted several subgroup analyses. Based on the hypothesis that frail patients would have lower utilization of life-sustaining therapy, greater withholding/withdrawal of life-support orders, and worse clinical outcomes than patients who were not frail, we compared treatments and outcomes in patients based on a Clinical Frailty Scale score of 5 or more versus less than 5. We also examined whether there were differences in treatments and outcomes of care based on the presence or absence of advance directives and family preferences for medical treatments at the EOL. The Kruskal-Wallis test (for three-group comparisons) or the Wilcoxon-Mann-Whitney test (for two-group comparisons) was used to compare continuous variables between groups. Categorical variables were compared between groups by the chi-square test. Continuous variables were described as means, SD, and the minimum and maximum values except for length of stay variables, which were described by their quartiles due to their positive skew. Categorical variables were described by counts and percentages. A multinomial logistic regression model was used to examine the independent association between various patient characteristics and the odds of dying within 7 days or surviving to hospital discharge, compared with dying in the hospital after remaining in the ICU for at least 7 days. All analyses were done using SAS Version 9.3 (SAS, Cary, NC).
Methods
We conducted a multicenter, prospective, observational cohort study with a nested cohort from September 2009 to February 2013 in 24 Canadian ICUs (for participating centers, see Appendix 1, Supplemental Digital Content 1, http://links.lww.com/CCM/B299). We report here an analysis of the ICU treatments and hospital outcomes in all patients (hospital cohort). Details of long-term outcomes of the nested cohort have been reported elsewhere (Heyland et al, unpublished observations, 2015). Institutional research ethics board approval was obtained from each center. Written informed consent was obtained from each patient's next of kin before enrollment.
In participating centers, all patients aged 80 or more were eligible for the hospital outcomes cohort; inclusion was based on research coordinator availability and resources to support the data collection. We obtained a waiver of informed consent to collect data that was readily available from the hospital charts. However, patients in the hospital were eligible for enrollment into a nested cohort study for which consent was required. An eligible family member was required for participation in the follow-up study. Patients were excluded from the nested cohort if they were imminently dying, were not expected to stay in ICU for more than 24 hours, had no permanent address, or were not Canadian residents (due to anticipated challenges in posthospital follow-up). An eligible family member consisted of a family member including partner, significant other, and/or close friend who: 1) provided the most care to the patient before hospital admission and was not paid to do so; 2) visited the patient at least once during the index ICU admission; and 3) was 18 years old or older. We also excluded patients from the nested cohort who did not have a French- or English-speaking family member.
Study Procedures
Baseline demographic and clinical data collected from the medical chart for all patients included: age, sex, ICU admission diagnosis, admission type (medical vs surgical), Acute Physiology and Chronic Health Evaluation II score, functional comorbidity index, and Charlson comorbidity index. We assessed sequential organ failure assessment scores at admission and daily while in ICU.
For patients who were enrolled in the nested follow-up, research coordinators met with the family member to collect additional data, including a measure of baseline frailty using the Clinical Frailty Scale. This score ranges from 1 (very fit) to 7 (very frail) and has been previously validated in the ICU. In addition, we documented whether the patient had a living will or advance directive about preferred medical care if he/she was seriously ill. We did not record the contents of the advance directive; just whether the patient had completed one. Finally, we asked the patient's family member which of the following three options they preferred for their loved one: 1) "Life Support. This option can include the use of a breathing machine as well as drugs and procedures to maintain body functions. Efforts are made to keep the patient as comfortable as possible." 2) "Comfort Care without Life Support. Comfort care aims to relieve suffering and preserve the dignity of the patient, without prolonging the dying process. There are usually fewer tests, fewer tubes, and no life support machines or monitors connected to the patient." 3) "I am unsure."
Use of life-sustaining treatment (mechanical ventilation, vasoactive drugs, and renal replacement therapy) and ICU and hospital length of stay were abstracted from the medical record in all patients. Medical orders for administration, withholding, or withdrawing of life-sustaining treatment before ICU admission and during ICU stay were also recorded. Withholding life-sustaining treatment was defined as no current receipt of any such treatment followed by a written order not to start or restart it. Withdrawing life-sustaining treatment was defined as current receipt of any such treatment followed by a written order to discontinue it in anticipation of death.
Statistical Analysis
Our primary outcome was a prolonged dying experience defined arbitrarily as staying in ICU for 7 days or longer and dying during the hospitalization. Baseline patient characteristics, treatments, and outcomes were compared between patients who 1) died within 7 days of ICU admission, 2) died at any point in hospital after staying in ICU for at least 7 days, and 3) survived to hospital discharge. In addition to the overall analyses, in the nested cohort, we conducted several subgroup analyses. Based on the hypothesis that frail patients would have lower utilization of life-sustaining therapy, greater withholding/withdrawal of life-support orders, and worse clinical outcomes than patients who were not frail, we compared treatments and outcomes in patients based on a Clinical Frailty Scale score of 5 or more versus less than 5. We also examined whether there were differences in treatments and outcomes of care based on the presence or absence of advance directives and family preferences for medical treatments at the EOL. The Kruskal-Wallis test (for three-group comparisons) or the Wilcoxon-Mann-Whitney test (for two-group comparisons) was used to compare continuous variables between groups. Categorical variables were compared between groups by the chi-square test. Continuous variables were described as means, SD, and the minimum and maximum values except for length of stay variables, which were described by their quartiles due to their positive skew. Categorical variables were described by counts and percentages. A multinomial logistic regression model was used to examine the independent association between various patient characteristics and the odds of dying within 7 days or surviving to hospital discharge, compared with dying in the hospital after remaining in the ICU for at least 7 days. All analyses were done using SAS Version 9.3 (SAS, Cary, NC).