Association between Goals of Care Designation orders and health care resource use among seriously ill older adults in acute care: a multicentre prospective cohort study

Madalene A. Earp; Konrad Fassbender; Seema King; Maureen Douglas; Patricia Biondo; Amanda Brisebois; Sara N. Davison; Winnie Sia; Eric Wasylenko; LeAnn Esau; Jessica Simon

doi:10.9778/cmajo.20210155

Article Figures & Tables

Figures

Tables

Download figure
Open in new tab
Download powerpoint
Figure 1:
Flow chart showing recruitment of participants. Note: DAD = Discharge Abstract Database.
Download figure
Open in new tab
Download powerpoint
Figure 2:
Ribbon plot showing study participants’ change in Goals of Care Designation (GCD) type over time, from pre–hospital admission to study enrolment, to last recorded at discharge, death or end of study. Colour corresponds to GCD type (resuscitative care GCD: dark green; medical care GCD: light green; comfort care GCD: orange; no GCD: red). Line thickness corresponds to the proportion of the sample.

Tables

Figures

View popup

Table 1:

Demographic and medical characteristics of study participants at enrolment

Characteristic	No. (%) of participants^*; GCD type at enrolment
Characteristic	All n = 475	Resuscitative GCD n = 177	Medical GCD n = 248	Comfort GCD n = 20	No GCD n = 30
Patient characteristics
Age at enrolment, yr, median (IQR)	83 (77–87)	81 (75–86)	85 (80–89)	79 (70–86)	80 (71–84)
Sex, female	249 (52.4)	87 (49.2)	131 (52.8)	9 (45.0)	22 (73.3)
Race and ethnicity, white^†	439 (92.4)	160 (90.4)	233 (94.0)	18 (90.0)	28 (93.3)
Educational attainment
Less than high school	159 (33.5)	52 (29.4)	88 (35.5)	4 (20.0)	15 (50.0)
High school	112 (23.6)	42 (23.7)	58 (23.4)	4 (20.0)	8 (26.7)
College (some or complete)	101 (21.3)	37 (20.9)	54 (21.8)	6 (30.0)	4 (13.3)
University (any level)	77 (16.2)	42 (23.7)	47 (19.0)	6 (30.0)	3 (10.0)
Missing	5 (1.1)	4 (2.3)	1 (0.4)	0 (0)	0 (0)
Residence before admission^‡
Home	354 (74.5)	138 (78.0)	176 (71.0)	14 (70.0)	26 (86.7)
Retirement residence	81 (17.1)	31 (17.5)	44 (17.7)	4 (20.0)	2 (6.7)
Long-term care	37 (7.8)	6 (3.4)	27 (10.9)	2 (10.0)	2 (6.7)
Other	3 (0.6)	2 (1.1)	1 (0.4)	0 (0)	0 (0)
Weighted Charlson Comorbidity Index score^§
Low (0–2)	294 (61.9)	124 (70.1)	146 (58.9)	4 (20.0)	20 (66.7)
Moderate–high (≥ 3)	181 (38.1)	53 (29.9)	102 (41.1)	16 (80.0)	10 (33.3)
Clinical Frailty Scale score^§
Well (1–3)	135 (28.4)	69 (39.0)	54 (21.8)	1 (5.0)	11 (36.7)
Mildly frail (4–6)	315 (66.3)	100 (56.5)	182 (73.4)	16 (80.0)	17 (56.7)
Severely frail (7–9)	25 (5.3)	8 (4.5)	12 (4.8)	3 (15.0)	2 (6.7)
Died during index admission	35 (7.4)	6 (3.4)	23 (9.3)	4 (20.0)	2 (6.7)
City^¶
Calgary	240 (50.5)	71 (40.1)	145 (58.5)	14 (70.0)	5 (16.7)
Edmonton	140 (29.5)	45 (25.4)	74 (29.8)	5 (25.0)	16 (53.3)
Lethbridge	95 (20.0)	56 (31.6)	29 (11.7)	1 (5.0)	9 (30.0)
Inclusion characteristics
Age ≥ 80 yr, no condition	283 (59.6)	112 (63.3)	150 (60.5)	15 (75.0)	15 (50.0)
“Surprise question”^§	4 (0.8)	1 (0.6)	2 (0.8)	0 (0)	1 (3.3)
Age ≥ 55 yr, with condition	188 (39.6)	64 (36.2)	96 (38.7)	14 (70.0)	14 (46.7)
Chronic obstructive lung disease	72 (15.2)	28 (15.8)	36 (14.5)	1 (5.0)	7 (23.3)
Cancer	67 (14.1)	13 (7.3)	39 (15.7)	13 (65.0)	2 (6.7)
Congestive heart failure	36 (7.6)	12 (6.8)	22 (8.9)	0 (0)	2 (6.7)
Renal failure	17 (3.6)	9 (5.1)	6 (2.4)	0 (0)	2 (6.7)
Cirrhosis	7 (1.5)	4 (2.3)	2 (0.8)	0 (0)	1 (3.3)

Note: GCD = Goals of Care Designation, IQR = interquartile range.
↵* Unless stated otherwise.
↵† Race and ethnicity as self-reported by participants. Other responses for self-reported race and ethnicity included Asian (n = 20), Aboriginal (n = 10) and Black (n = 5). Data missing for 1 participant.
↵‡ Long-term care description includes nursing homes and residential care as 1 category; “other” includes rehabilitation facility (n = 1), hospital (n = 1) and other not specified (n = 1).
↵§ For weighted Charlson Comorbidity Index score, see references (30)^, (31) and Methods, for Clinical Frailty Scale score see reference (29) and Methods, for “surprise question” see references (15)^, (16) and Methods.
↵¶ Calgary sampling at Foothills Medical Centre (units 32, 46, 47, 61) n = 93, Peter Lougheed Centre (units 38, 39, 42, 43, 49) n = 53, Rockyview General Hospital (units 72, 73, 93, 94) n = 94; Edmonton sampling at Grey Nuns Community Hospital (units 45, 51, 53, 54) n = 76, Royal Alexandra Hospital (units 41, 44, 51–53, 5E, 5W, 6E, SW) n = 24, University of Alberta Hospital (units 5A4, 5D3, 5D4, 5F2, 5F3, 500, 5000, 50 000) n = 40; Lethbridge sampling at Chinook Regional Hospital (units 4B, 4C) n = 95.

View popup

Table 2:

Health care resource use during hospitalization, by GCD type

Outcome	GCD type at enrolment								p value^*
Outcome	Resuscitative care n = 177		Medical care n = 248		Comfort care n = 20		No GCD n = 30		p value^*
Continuous	Median (IQR)
All days LOS	7 (5–12)		9 (6–20)		10 (6.8–16)		6.5 (4.2–10)		< 0.001
Resource Intensity Weight	1.0 (0.7–1.6)		1.3 (0.8–2.5)		1.5 (1.0–1.9)		1.1 (0.7–1.6)		0.002
Average cost per stay, $^†	7893.00		10 260.90		11 839.50		8682.30		–
Dichotomous	No. (%)	95% CI	No. (%)	95% CI	No. (%)	95% CI	No. (%)	95% CI
Admission to ICU, yes	3 (1.7)	0–5	0 (0)	0–1	0 (0)	0–1	1 (3.3)	0–17	0.08
Inpatient palliative care referral, yes	9 (5.1)	2–9	49 (19.8)	15–25	18 (90.0)	68–99	0 (0)	0–11	< 0.001
Flagged intervention, any^§	38 (21.5)	16–28	47 (19.0)	14–24	7 (35.0)	15–59	3 (10.0)	2–27	0.2
Biopsy	13 (7.3)	4–12	12 (4.8)	3–8	1 (5.0)	0–25	2 (6.7)	0–22	0.8
Endoscopy	13 (7.3)	4–12	10 (4.0)	2–7	1 (5.0)	0–25	0 (0)	0–11	0.2
Pleurocentesis^¶	6 (3.4)	1–7	12 (4.8)	3–8	2 (10.0)	1–32	0 (0)	0–11	0.3
Radiotherapy^¶	3 (1.7)	0–5	7 (2.8)	1–6	2 (10.0)	1–32	0 (0)	0–11	0.1
Dialysis	5 (2.8)	1–6	4 (1.6)	0–4	0 (0)	0–17	1 (3.3)	0–17	0.7
Paracentesis^¶	4 (2.3)	1–6	4 (1.6)	0–4	2 (10.0)	1–32	0 (0)	0–11	0.07
Vascular access device	5 (2.8)	1–6	5 (2.0)	1–5	0 (0)	0–17	0 (0)	0–11	0.7

Note: CI = confidence interval, GCD = Goals of Care Designation, ICU = intensive care unit, IQR = interquartile range, LOS = length of stay.
↵* p value for LOS and Resource Intensity Weight from Krustal–Wallis test given 4 GCD categories (resuscitative, medical, comfort and no GCD); p values for all dichotomous variables from Fisher exact test given same 4 GCD categories. Tests implemented in R version 4.0.0.
↵† Calculated as: Resource Intensity Weight × Alberta’s cost of standard hospital stay for 2017–2018 ($7893).
‡ Does not include intensive palliative care unit admission. Average hours in ICU for resuscitative GCD = 1.97; no GCD = 0.53.
↵§ Flagged interventions that occurred in ≤ 1 participant not shown. Never occurred: cell saver, invasive ventilation ≥ 96 h, invasive ventilation < 96 h, parenteral nutrition, tracheostomy. Occurred once: chemotherapy (resuscitative GCD), cardioversion (resuscitative GCD), heart resuscitation (resuscitative GCD) and feeding tube (resuscitative GCD).
↵¶ These are flagged interventions that may be used to provide symptom relief or comfort care.

View popup

Table 3:

Association between GCD at enrolment and health care resource use during hospitalization

Outcome	Unadjusted rate ratio (95% CI)	Adjusted rate ratio (95% CI)
Continuous
All days LOS^*
Resuscitative care	Ref.	Ref.
Medical care	1.64 (1.26–2.13)	1.42 (1.10–1.83)
Comfort care	1.73 (0.92–3.26)	1.26 (0.68–2.36)
No GCD	0.75 (0.44–1.27)	0.76 (0.45–1.27)
Resource Intensity Weight^†
Resuscitative care	Ref.	Ref.
Medical care	1.20 (1.07–1.34)	1.14 (1.02–1.28)
Comfort care	1.55 (1.19–2.03)	1.13 (0.87–1.48)
No GCD	1.00 (0.80–1.24)	0.99 (0.80–1.23)
Dichotomous^‡	Unadjusted relative risk (95% CI)	Adjusted relative risk (95% CI)
Inpatient palliative care referral, yes
Resuscitative care	Ref.	Ref.
Medical care	3.88 (1.96–7.7)	3.58 (1.75–7.33)
Comfort care	17.63 (9.21–34.01)	9.32 (4.32–20.08)
No GCD^§	–	–
Flagged intervention, any
Resuscitative care	Ref.	Ref.
Medical care	0.98 (0.92–1.04)	0.98 (0.48–2.02)
Comfort care	1.11 (0.95–1.31)	1.06 (0.49–2.28)
No GCD	0.91 (0.81–1.01)	0.90 (0.39–2.07)

Note: CI = confidence interval, GCD = Goals of Care Designation, LOS = length of stay, Ref. = reference category.
↵* Truncated negative binomial regression adjusted for age at enrolment, sex, race and ethnicity (white v. other), Clinical Frailty Scale score (well, mildly frail, severely frail), weighted Charlson Comorbidity Index score (low, moderate–high) and city. Truncated at 2 given patients were ineligible during the first 48 hours of an admission.
↵† γ-log link regression adjusted for age at enrolment, sex, race and ethnicity (white v. other), Clinical Frailty Scale score (well, mildly frail, severely frail), weighted Charlson Comorbidity Index score (low, moderate–high) and city.
↵‡ Modified Poisson regression model adjusted for age at enrolment, sex, race and ethnicity (white v. other), Clinical Frailty Scale score (well, mildly frail, severely frail), weighted Charlson Comorbidity Index score (low, moderate–high) and city.
↵§ Had no inpatient palliative care referral, so omitted from analysis.

View popup

Table 4:

Association between GCD at discharge and readmission in 30 days

Outcome^*	GCD type at discharge from acute care
Outcome^*	No./total (%)	Unadjusted relative risk (95% CI)	Adjusted relative risk (95% CI)
Hospital readmission from any discharge location in 30 days^†
Resuscitative care	26/168 (15.5)	Ref.	Ref.
Medical care	47/219 (21.5)	1.05 (0.99–1.12)	1.05 (0.97–1.20)
Comfort care	2/33 (6.1)	0.92 (0.84–1.01)	1.00 (0.85–1.19)
No GCD	1/17 (5.9)	0.92 (0.82–1.00)	0.90 (0.80–1.01)
Hospital readmission from home only in 30 days^‡
Resuscitative care	23/134 (17.2)	Ref.	Ref.
Medical care	39/154 (25.3)	1.07 (0.99–1.16)	1.06 (0.98–1.16)
Comfort care	1/5 (20.0)	1.02 (0.76–1.38)	1.01 (0.74–1.36)
No GCD	1/16 (6.3)	0.91 (0.80–1.03)	0.91 (0.80–1.03)

Note: CI = confidence interval, GCD = Goals of Care Designation, Ref. = reference category.
↵* Compared with participants with a resuscitative GCD.
↵† Modified Poisson regression model adjusted for discharge location, age at enrolment, sex, race and ethnicity (white v. other), Clinical Frailty Scale score (well, mildly frailty, severely frail), weighted Charlson Comorbidity Index score (low, moderate–high), and city.
↵‡ Modified Poisson regression model adjusted for age at enrolment, sex, race and ethnicity (white v. other), Clinical Frailty Scale score (well, mildly frail, severely frail), weighted Charlson Comorbidity Index score (low, moderate–high) and city.