Table 3.

Post-Discharge Exposures Associated with Follow-up Hospitalization and Episode of Care Costs After Acute Respiratory Distress Syndrome

Post-Discharge Exposurea	Any Hospitalization		Episode of Care Cost
	Odds Ratio (95% CI)	P Valueb	Relative median (95% CI)	P Valuec
Physical assessment
Grip strengthd, per 6 kg	0.85 (0.73, 1.00)	0.047	0.77 (0.69, 0.87)	<0.001
Grip strength % predictedd, per 12%	0.86 (0.77, 0.97)	0.013	0.89 (0.81, 0.97)	0.006
6 minute walk distancee, per 30m	0.94 (0.89, 1.00)	0.045	0.93 (0.89, 0.97)	0.001
6 minute walk % predictede, per 5%	0.94 (0.88, 0.99)	0.025	0.93 (0.89, 0.97)	<0.001
ADL number of dependencies	0.97 (0.95, 1.12)	0.717	1.07 (0.96, 1.21)	0.217
IADL number of dependencies	1.05 (0.96, 1.14)	0.268	1.08 (1.02, 1.16)	0.012
Physical impairmentf	1.30 (0.89, 1.90)	0.174	1.19 (0.87, 1.63)	0.270
Psychiatric symptoms
Depressiong	1.22 (0.82, 1.82)	0.331	0.90 (0.65, 1.25)	0.532
HADS depression subscalef, per 1.5	1.04 (0.96, 1.12)	0.314	1.05 (0.98, 1.11)	0.170
Anxietyg	0.98 (0.67, 1.45)	0.937	0.95 (0.68, 1.31)	0.746
HADS anxiety subscalef, per 1.5	1.00 (0.93, 1.08)	0.952	1.03 (0.97, 1.10)	0.284
PTSDd,h	1.27 (0.77, 2.11)	0.348	1.17 (0.80, 1.70)	0.412
IES-R scored per 0. 4	1.03 (0.93, 1.14)	0.587	1.02 (0.95, 1.11)	0.553
Health-Related Quality of life
EQ-5D VASd, per 10	0.94 (0.86, 1.03)	0.166	0.97 (0.91, 1.04)	0.442
EQ-5D utility scoree, per 0.07	0.92 (0.87, 0.98)	0.008	0.95 (0.90, 0.99)	0.013
SF-36 PCSe, per 5	0.85 (0.78, 0.94)	0.001	0.92 (0.86, 0.98)	0.016
SF-36 MCSe, per 5	1.01 (0.94, 1.08)	0.792	0.99 (0.94, 1.05)	0.820

CI = confidence interval, kg = kilogram, m = meter, ADL = activities of daily living, IADL = instrumental activities of daily living, HADS = Hospital Anxiety and Depression Scale, PTSD = post-traumatic stress disorder, IES-R = impact of event scale-revised, EQ-5D = EuroQol 5-dimension questionnaire, VAS = visual analogue scale, SF-36 = short form 36 questionnaire, PCS = physical component score, MCS = mental component score.

^aFor each of the logistic and linear models, each post-discharge exposure was added separately to the same base model that adjusted for sex, informal caregiver status, insurance status, Charlson comorbidity index, intensive care unit length of stay, and time (i.e. year of follow-up).

^bP values were calculated using multivariable logistic regression analysis of patients with any versus no hospitalization during each follow-up year. Hosmer-Lemeshow test confirmed model fit.

^cP values were calculated using multivariable linear regression analysis of log-transformed yearly episode of care costs associated with each hospitalization, including the costs of skilled nursing and rehabilitation facility costs which were associated with hospitalizations. (17)

^dGrip strength, IES-R score, and EQ-5D-VAS are all scaled by one-half of the standard deviation (30) within this study cohort.

^e6MWD, HADS, EQ-5D-utility score, SF-36 PCS and MCS, and were scaled by the reported minimally important difference (26–29).

^fPhysical impairment was defined by ≥2 dependencies in independent activities of daily living.

^gAnxiety and depression symptoms defined by HADS scores ≥8.

^hPTSD defined by IES-R score ≥1.6.