Longterm Outcomes of the Very Elderly Admitted to

Long-term Outcomes of the Very Elderly Admitted to ICUs in Canada: A Quality Finish? Daren K. Heyland MD, MSc Queen’s University and Kingston General Hospital Kingston Ontario www. thecarenet. ca

Background • Globally, population is ageing • In US, the number of persons aged 85 years or older is likely to grow from about 4 million in 2000 to 19 million by 2050 • In Canada – Currently 55 -64 fast growing age group – >80 second fast growing age group (25% increase from 2001 -2005) – Rate of growth projected to continue till 2031 when seniors would account for 25% of population (currently 13%) – As the population ages, the proportion of patients with advanced medical diseases will rise and our health care system needs to adapt to this changing demographic

Current ICU Utilization of Very Elderly in Ontario 18% 22% www. thecarenet. ca

Biomedicalization of Aging • Society views ageing negatively, as a disease for which medical intervention is normalized, necessary, and appropriate. • The more high tech medicine one consumes; the more one forestalls aging. • In this paradigm, use of life sustaining technology becomes the norm • From Family’s perspective, caring is associated with everything being done • Consequence is that we have a crisis where demand for critical resources outstrips supply and problem will only get worse

Background Quality EOL Care Scare ICU Resources Aging of Society (Very) Elderly in the ICU? www. thecarenet. ca

A Quality Finish? CMAJ 2006; 174: 627 www. thecarenet. ca

What Matters the Most to Quality of End of Life Care in Canada N= 440 Results from Patient’s Perspective Areas of Greatest “Importance” % “Extremely Important” To have trust and confidence in the Doctor looking after you 55. 8 Not to be kept alive on life support when there is little hope for a meaningful recovery 55. 7 That information about your disease be communicated to you in a honest manner 44. 1 To complete things and prepare for life’s end 43. 9 To have an adequate plan of care and services available to look after you at home upon discharge 41. 8 To not be a physical or emotional burden on your family 41. 8 Good Communication and Decision-Making? www. thecarenet. ca Heyland CMAJ 2006; 174: 627

Improving EOL communication and decision making Greatest Improvements in EOL Care Improving EOL communication and decision making www. thecarenet. ca

Failure to Engage Hospitalized Elderly Patients and Their Families • Multicenter survey of 283 80+ on hospital wards • Majority had thought of EOL wishes and could • • • express preference for treatment at EOL Less than 1/3 had spoken to health care professional Fewer than 20% acknowledged a prognostic disclosure Expressed preferences and documents ‘goals of care’ only agreed 1/3 time www. thecarenet. ca Heyland JAMA Int Med 2013

Consequences of Intensification of Care at the End of Life • 7 academic centers from US • 332 patients with advanced cancer who died and their • • • family caregivers Patients asked, “have you and your doctor discussed any particular wishes you have about the care you would want to receive if you were dying? ” 123 (37%) reported have end of life discussion at baseline median follow up 4. 4 months after baseline www. thecarenet. ca Wright JAMA 2008; 300: 1665

End of Life Discussions… …Associated with decreased intensity of care www. thecarenet. ca Wright JAMA 2008; 300: 1665

plau • Background: – 30% of medical care is attributable to 5% of those who die in a year – About 1/3 of annual expenditures of those who die occurs in the last month – Most of these final costs are secondary to aggressive medical care in the last 30 days. • Patients with no EOL conversation much more likely to • • die in ICU, less likely to go to hospice Last week of life medical expenses $2917 vs $1876 for those who did have a EOL conversation Based on number of cancer deaths each year, if could increase ACP to 50% levels, would result in $76 million savings/year. www. thecarenet. ca Zhang Arch Intern Med 2009; 169: 480

Summary • Very elderly- it is plausible that poor communication and decision-making leads to overutilization of ICU resources and poor quality EOL care. • To the extent that admission to ICU for an elderly patient requires a decision to be made, let’s improve clinical decision making!

Purpose of Information Exchange Prognosis + Values Good Decision What do we know about the outcomes of the very elderly critically ill patient? www. thecarenet. ca

REALISTIC-80 What do we know about outcomes of Elderly in ICU • European and US studies: – ICU mortality 30 -35% mortality – 12 month mortality 60 -70% mortality – Severity of illness strongest predictor of short term survival – Comorbidities strongest predictor of long term survival – Significant comorbidities plus prolonged ICU stay=<5% survival • Limited data on functional outcomes/QOL – QOL studies in selected survivors – Of survivors, functional status or QOL seems ‘reasonable’ – Most studies old, single centered, using non-validated instruments of functional status • No studies have comprehensively evaluated the determinants of long-term quality of life or functional recovery after critical illness in very elderly persons.

Information Most Important to Patients Facing a Life-threatening Illness Ø Most Important Ø chances of surviving Ø resultant health state More important for very elderly Ø Moderate Importance Ø Impact on family’s lives Ø Least Important Ø Length of hospital stay, Ø probability of institutionalization Ø amount of pain Ø ICUs, ventilators etc. Heyland Chest 2006; 130: 419 and Lloyd CCM 200432: 649

Realities, Expectations and Attitudes to Life Support Technologies in Intensive Care for Octogenarians The REALISTIC 80 study Funded by CIHR Conducted under the auspices of the CCCTG and CARENET www. thecarenet. ca

REALISTIC-80 Overall Design All 80+ admitted to participating ICU >24 hrs and had a family member present 610 Enrolled ICU 80+ Patients and their families 3, 6, 9, 12 Month Outcomes Inclusion criteria: - >24 hrs in ICU - Family present Heyland Crit Care Med 2015; Palliative Med 2015; Intensive Care Med 2015 (in press)

REALISTIC-80 Research Questions-Patient • Primary 1) What are the 12 month survival and HRQOL of patients admitted to ICU who are 80+ years old? • Secondary 2) Which patient characteristics are associated with recovery from critical illness at 12 months? We defined ‘recovery’ from critical illness as being alive with SF-36 physical function score of at least 10 points and not 10 or more points below baseline at 12 months.

REALISTIC-80 Research Questions-Family 3) For non-surviving patients, what are the processes of care (descriptive)? 4) For non-surviving patients, what is the family satisfaction with EOL care, as measured using Family Satisfaction with ICU Care 24 (in ICU death) and the CANHELP Satisfaction instrument (in hospital)? 5) What are the values that influence decisions about goals of care for this patient population? 6) What is the quality of decision making about the goals of care for an 80+ patient?

Patient Characteristics Age Sex (Male) Baseline APACHE II Baseline SOFA score Admission types Medical Surgical elective Surgical emergency Primary ICU diagnosis Cardiovascular/vascular Respiratory Gastrointestinal Sepsis Other Baseline Physical Function score Charlson Co-morbidity Index Functional Co-morbidity Index IQCODE Frailty Index www. thecarenet. ca All patients (n=610) 84. 4± 3. 4 (79. 9 -99. 4) 338 (55. 4%) 21. 6± 6. 9 (7. 0 -49. 0) 5. 4± 3. 2 (0. 0 -15. 0) 377 (61. 8%) 83 (13. 6%) 150 (24. 6%) 143 (23. 4%) 157 (25. 7%) 110 (18. 0%) 72 (11. 8%) 128 (21. 0%) 40. 3± 29. 9 (0. 0 -100. 0) 2. 1± 1. 9 (0. 0 -11. 0) 1. 9± 1. 4 (0. 0 -6. 0) 3. 3± 0. 5 (1. 0 -5. 0) 0. 3± 0. 1 (0. 0 -0. 7)

Family Member Characteristics Family member age * Family member sex male female The patient is my… Father/Mother Husband/Wife Brother/Sister Other Missing Are you the person who provides the most care for this person? Yes No Missing Are you the legal substitute decision maker for the patient? Yes No Unsure www. thecarenet. ca All Family Members (n=535) 61. 4± 13. 6 (18. 0 -100. 0) 162 (30. 3%) 373 (69. 7%) 337 (63. 0%) 153 (28. 6%) 10 (1. 9%) 34 (6. 4%) 1 (0. 2%) 430 (80. 4%) 104 (19. 4%) 1 (0. 2%) 395 (73. 8%) 113 (21. 1%) 26 (4. 9%)

Clinical Outcomes All patients (n=610) Proportion of patients undergoing invasive mechanical 439 (72. 0%) ventilation Average duration of invasive mechanical ventilation 4. 0[2. 0 to 10. 0] (1. 0, 116. 0) (median, IQR, range) Index ICU LOS (median, IQR, range) 6. 0[ 4. 0 to 11. 0] (2. 0, 96. 0) Total Hospital LOS (median, IQR, range) 21. 1[11. 3 to 39. 6] (0. 4, 201. 8) Hospital LOS * Percentage of patients have a stay in hospital prior to ICU 554 (90. 8%) ICU mortality 85 (13. 9%) Hospital mortality 158 (25. 9%) Discharged from Hospital: 452 (74. 1%) Ward in another hospital 132 (21. 6%) ICU in another hospital 11 (1. 8%) Long term care facility 81 (13. 3%) Home 197 (32. 3%) Rehab 25 (4. 1%) Palliative Care 1 (0. 2%) Other 5 (0. 8%) Proportion of patient at home prior to hospitalization 179 (34. 0%) discharged to home www. thecarenet. ca

Kaplan-Meier Survival Curves of Study Population Compared to Age and Sex Matched Community Control Population. www. thecarenet. ca

SF-36 Scores Among Survivors: Physical Functioning Domain www. thecarenet. ca Heyland ICM 2015 (in press)

Recovery to Baseline Physical Function After Critical Illness Among Patients Aged 80 Years or Older www. thecarenet. ca Heyland ICM 2015 (in press)

Logistic Regression Model Predicting Physical Recovery 12 months After ICU Admission Variables Multivariable Predictor Model OR (95% CI) P-value Age (per 5 years) 0. 73 (0. 57, 0. 93) 0. 01 Sex (Male vs. Female) APACHE II score (per 10 points) Baseline SOFA score (per 5 points) Admission type (Medical vs. Surgical) Surgical elective vs. Medical Surgical emergency vs. Medical Primary ICU diagnosis CABG/Valve vs. Cardiovascular/vascular Gastrointestinal vs. Cardiovascular/vascular Neurologic vs. Cardiovascular/vascular Other vs. Cardiovascular/vascular Respiratory vs. Cardiovascular/vascular Sepsis vs. Cardiovascular/vascular Stroke vs. Cardiovascular/vascular Trauma vs. Cardiovascular/vascular Baseline PF score (per 50 points) Charlson Comorbidity Index (per 2 units) IQCODE at baseline (per 0. 5 point) Frailty Index (per 0. 2 point) 0. 88 (0. 57, 1. 34) 0. 50 (0. 27, 0. 96) 0. 82 (0. 57, 1. 19) 1. 71 (0. 79, 3. 67) 1. 83 (0. 77, 4. 34) 4. 21 (1. 96, 9. 03) 1. 09 (0. 53, 2. 22) 1. 30 (0. 53, 3. 18) 0. 77 (0. 26, 2. 25) 1. 24 (0. 48, 3. 24) 1. 38 (0. 54, 3. 51) 0. 11 (0. 01, 0. 91) 0. 47 (0. 15, 1. 47) 0. 32 (0. 22, 0. 45) 0. 76 (0. 59, 0. 98) 0. 96 (0. 71, 1. 30) 0. 32 (0. 19, 0. 56) 0. 55 0. 04 0. 29 0. 32 0. 0001 <0. 0001 0. 03 0. 77 <0. 0001 Heyland ICM 2015 (in press)

Clinical Prediction Rule Table 1: Calculation of recovery scale Characteristics Recovery Points Age<90 1 Apache 2 <20 1 Surgical Admission (elective or emergency) 1 Frailty Index <. 15 2 Frailty Index . 15 to <. 25 1 Baseline SF 36 PF domain <35 2 Baseline SF 36 PF domain 35 to <65 1 Sum points for possible range of 0 to 7. www. thecarenet. ca

REALISTIC-80 Research Questions-Family 3) For non-surviving patients, what are the processes of care (descriptive)? 4) For non-surviving patients, what is the family satisfaction with EOL care, as measured using Family Satisfaction with ICU Care 24 (in ICU death) and the CANHELP Satisfaction instrument (in hospital)? 5) What are the values that influence decisions about goals of care for this patient population? 6) What is the quality of decision making about the goals of care for an 80+ patient?

The Very Elderly Admitted to Intensive Care Unit: A quality finish ? • Of enrolled patients, 240 (39%) remained in ICU for 7 • days or more; of these, 99 (41%) died in hospital. Neither frailty nor advance directives had significant impact on processes of care (use or withhold/withdraw of life-sustaining treatments). www. thecarenet. ca Heyland CCM 2015

Additional Family Member Perspectives • On average, interviews occurred 3 days after ICU • admission Family members reported that the “patient be comfortable and suffer as little as possible” was the most important value and “the belief that life should be preserved at all costs” was the least important value considered in making their treatment decisions. www. thecarenet. ca Heyland Palliative Medicine 2015

Family Member Preferences • 57% reported that the doctor talked to them about being able to choose between treatment options for your family member in ICU. • 30. 7% of family members had decisional conflict related to their treatment preference. • Decisional conflict was lower in family members who had talked to a doctor than those who had not. Heyland Palliative Medicine 2015 www. thecarenet. ca

Additional Family Member Perspectives • If family members’ stated preferred comfort measures – 83. 7% of related patients received life-sustaining treatments; – 20. 2% received one or more for more than 7 days – the time from ICU admission to death was on average 10. 0 days amongst non-survivors. • Among non-survivors, time from ICU admission to death was longest in patients whose family members were ‘unsure’ of their treatment preferences (16. 0 days vs. 10. 0 days in comfort group vs. 12. 0 days in lifesustaining treatments group; p=0. 06). www. thecarenet. ca Heyland Palliative Medicine 2015

Family Satisfaction with Critical Care (FS-ICU 24) www. thecarenet. ca

Conclusions (1) Ø If admitted to ICU Ø Stay in ICU for a week or more Ø Remain in hospital for a month or more Ø 1/3 will die in hospital after a prolonged dying experience Ø 1/3 will be discharged home Ø 1/3 will be discharged to alternative location Ø At 1 year Ø Half patients will be dead Ø 25% will have a good outcome (recovery to baseline) Ø Narrative can be adapted to individuals with clinical prediction tool

Conclusions (2) Ø Baseline PF and Frailty Index are significant risk factors and consideration should be given to routinely measuring them in older patients admitted to ICU in addition to traditional measure describing the acute illness. Ø There is incongruity between family members’ values and preferences for end of life care of their very elderly relatives, and the actual care received. Ø Deficiencies in communication and decision-making may be associated with non-beneficial and prolonged use of life-sustaining treatments in very elderly critically ill patients, many of whom ultimately die.

Where do we go from here? www. thecarenet. ca

Effectivenss of Multi-faceted Decision Support Intervention Ø Elements to include Ø Information leaflet on ‘how decisions are made’ and ‘your role’ and ‘how others cope’ Ø Systematically eliciting patient/family values, preferences (including ACP/AD) Ø Systematically offering non-curative, palliative care as a treatment option Ø Systematically obtain information on key determinants to recovery (baseline PF, Frailty, comorbidities, etc. ) Provide this information to clinical team early in stay

www. myicuguide. com

A Quality Finsih? More needs to be done urgently to improve EOL care for very elderly (in Canada)! www. thecarenet. ca