Background: Good symptom control is an essential factor of care at end-of-life (EOL).1 Patients with life-limiting illness or those actively dying, often experience substantial pain and/or discomfort, whether from their admitting diagnosis, procedures done in the intensive care unit (ICU)2, or related with a multiplicity of medical conditions1. Patient self-report is gold standard, however, ICU patients over time, may be unable to self-report, due to declining physical or cognitive function.3 Thus, validating the congruence of proxy perceived assessments related to patient symptom burden is important, as often treatment decisions may be made by family/other proxy decision makers.4-6
Methods: Prospective, correlational design was used to study the reported symptom burden of ICU patients' compared to proxy perception of patient symptoms.
Study was conducted in a 55-bed ICU, 559-bed, academic medical center in western United States. A convenience sample of 80 patients at high-risk of dying and 53 proxy responders were enrolled, based on following criteria: 1) >18 years older; 2) speak, read/understand English/Spanish; 3) able to complete self-report instrument written or alternative method (e.g. if questionnaire read to patient by researcher; 4) normal mental status, assessed by the Richmond Agitation-Sedation Scale (RASS)7 and the Confusion Assessment Method for the ICU (CAMU-ICU)8 and; 5) probability factor of 60-80% risk for hospital death. Exclusion: 1) diagnosis of cancer and receiving treatments; 2) principal diagnosis of dementia with altered cognitive status; or 3) altered level of consciousness.
Proxies were defined as individuals who provided the majority of emotional, financial, and physical support to the patient prior to ICU admission. No legal relation or cohabitation with the patient was required. Eligibility were: 1) non-professional, non-paid caregiver; 2) age ≥ 18 years; 3) reliable telephone access; and 4) able to read and speak English/Spanish.
Patient variables were recorded on a demographic Log- age, sex, ethnicity of patient; insurance, diagnosis, admission location, specific treatments (mechanical ventilation; medications, particularly analgesics and sedatives, and comorbidities. The APACHE III Acute Physiologic Score (APS)9 and Predicted Mortality (r) for hospital death, was from the APACHE III data system.
Study instruments included: RASS, used to assess the level of agitation-sedation in patients prior to obtaining consent and conducting symptom measurement. Scores on the 10-point observational scale ranged from −5 (unarousable) to +4 (combative). Patients were judged able to provide informed consent and symptom measurement data if RASS score was 0 (alert and calm); Confusion Assessment Measure for the ICU (CAM-ICU) used to assess delirium prior to symptom measurement. CAM-ICU assesses four features of cognitive domain including: acute onset or fluctuating mental status, inattention, disorganized thinking, altered level of consciousness. Scores on four domains determined if delirium was present (CAM-ICU positive) or absent (CAM-ICU negative).
To assess patient symptom burden, the Modified Condensed Memorial Symptom Assessment Scale (CMSAS)10 was used, which is a shortened version of Memorial Symptom Assessment Sale (MSAS), a valid cancer tool to assesses multiple dimensions (severity, frequency, and distress) of 32 physical and emotional symptoms. The condensed version of CMSAS rates eleven physical symptoms in terms of distress from 0 (none) to 4 (very much), and three emotional symptoms in terms of frequency from 0 (not present) to 4 (almost constant). Proxies also completed the CMSAS- Proxy Respondent Symptom Survey, which was modified to state “Which of these symptoms bothered your loved ones in the last seven days” and “Which of these symptoms bothered your loved ones in the past two days”. Data points were day-1 and day-3 of the patients ICU stay. Proxies were asked to state their relationship to patient; age; ethnicity; and if they were primary care providers for patient. All study instruments (patient/proxy), were available in Spanish/English.
Procedures: Human subject approval was obtained from institution’s Institutional Review Board (IRB). Patients were screened daily using Acute Physiology, Age, Chronic Health Evaluation (APACHE III) score, and mortality risk (r) >.60% in first 24-hours for eligibility. APACHE III Scores and CMSAS data was recorded on Day-1-Day-3 of patient’s ICU stay. Patient informed consent was obtained if RASS score was 0 (alert and calm). If patient unable to give informed consent, proxy consent was obtained (n=44, 55%).
Statistical Analysis -Data was analyzed using IBM-SPSS v. 20.0 (SPSS, Inc.; Chicago, IL, USA). Descriptive statistics were used to analyze demographic and study variables. Chi-square test and Mann-Whitney U tests compared prevalence and severity scores of the four most prevalent symptoms in patients. Spearman’s rank correlation was used to explore the relationships between four most prevalent physical symptoms in patients and proxy responses. Crosstabs with the Pearson Chi-Square and Cohen’s Kappa statistics were computed to assess the concordance between patients and family members on ratings of CMSAS symptom presence. Statistical significance was set at α =0.05, two-tailed.
Results: The Condensed Memorial Symptom Assessment Scale (CMSAS) was used to gain patient/proxy report of symptoms on day-1 and day-3 after admission. N=80 patients who were able to respond to symptom assessments on day-1 while only 66 (80%) completed study survey on day-3. Reasons for attrition: patient expired (10%), patient discharged (5%), or patient lacked capacity (5%), with deterioration of critical illness. Overall, there was no significant difference in age, gender, ethnicity, and ICU admission source among those who did or did not complete day-3 symptom assessments.
Study Characteristics: Patient’s mean age was 70.5 ±; and primarily male (62.5%). Racial/ethnic composition was diverse with 48.8% White, 22.5% Black, 15% Hispanic, 11.3% Asian and 2.5% other. Participants were predominately English speaking (81.3%), yet 12.5% of the participants completed the symptom survey(s) in Spanish. Among the cohort, the first-ICU-day mean (%) predicted hospital mortality rate was 69%, and mean first-ICU-day APACHE III score was 170± 23. The mean ICU length of stay (LOS) was 10, with a range of 1-49 days and mean hospital LOS was 15, (range = 1.5-73) days. Following discharge from the ICU approximately 16% of the patients had to be readmitted to the ICU from the discharge unit for further treatment. The observed ICU and hospital mortality rates were 19% and 2% respectively. 30% of the patients who died in the ICU, resulted from the decision to withhold or withdraw life-sustaining treatments while in the ICU. Of the 63 patients who survived to hospital discharge, 16 (25%) died within three months following discharge.
Proxy Reporters - Mean age was 59 years (25-92) and (69.8%) were female. Racial/ethnicity was: White (43.4%), Black (17%), Hispanic (20.8%), Asian (15.1%) and other (3.8%). Relationship of proxies to patient included: spouse (43.4%), patient’s child (41.5%), patient’s sibling (5.7%), patient’s parent (3.8%) and significant partner (5.7%). Three quarters (75%) were English speaking, 9.4% Spanish speaking and 13.2% other languages. Approximately 70% identified themselves as the primary caregivers.
Symptom Burden -Majority of patients were symptomatic (98%), reporting an average of 10.23 symptoms. Most common symptoms reported on Day-1, were lack of energy (fatigue) and difficulty concentrating, with a mean symptom distress score of 2.96 (SD = 0.70) and 2.79 (SD = 0.84), scored on a scale of 1 = low symptom distress to 4 = high symptom distress, respectively. The CMSAS Total Distress Score was 2.24 (SD=0.66). The Physiological Symptom Distress Subscale (CMSAS-PHYS) score was 2.19 (SD=0.71). Approximately 97.9% of patients reported psychologic symptoms (sadness, worry, nervousness) with a mean symptom distress score of 2.45 (SD=0.66), measured by the Psychological Symptom Distress Subscale (CMSAS-PSYCH). On Day-3, 65 of the patients were still in the ICU. The most prevalent symptom reported was difficulty sleeping (90.8%), with a medium intensity distress score of 3.79 (SD=1.06). Eighty percent of patients reported additional symptoms: lack of energy, lack of appetite, pain, dry mouth, feeling drowsy, shortness of breath, and difficulty concentrating, with a moderate intensity mean score of 3.42.
Overall distress increased among all symptoms, as measured by the CMSAS-Total Distress Score of 3.17 (SD=0.44), and the two distress subscales: CMSAS-PHYS mean score of 3.07 (SD=0.46) and CMSAS-PSYCH means score of 3.46 (SD=0.52). Hospital mortality was 17 (21%) during initial hospitalization and 16 (25%) at 3-months follow-up. Family members correctly estimated the presence and absence of symptoms 85.5% of the time, yet rated the patients' physiologic symptoms higher than psychological distress.
Conclusions: This study identified that ICU patients near death, experience a significant burden of multiple symptoms, yet receive limited treatment for overwhelming symptom distress. A need for widespread symptom management strategies with proven effectiveness is indicated. Data also confirmed that proxy reporter’s perception of patient symptom burden can a reliable alternative, and should be used when patients can no longer self-report. Further research is needed to test new evidence-based interventions to serve as practice standards in the delivery of consistent, high quality care for all dying patients to minimize unnecessary suffering.