The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Early Active Mobilization during Mechanical

Ventilation in the ICU
The TEAM Study Investigators and the ANZICS Clinical Trials Group*​​

A BS T R AC T

BACKGROUND
Intensive care unit (ICU)–acquired weakness often develops in patients who are The members of the writing committee
undergoing invasive mechanical ventilation. Early active mobilization may mitigate (Carol L. Hodgson, Ph.D., Michael Bailey,
Ph.D., Rinaldo Bellomo, Ph.D., Kathy Brick-
ICU-acquired weakness, increase survival, and reduce disability. ell, R.G.N., Tessa Broadley, B.Biomed.Sci.,
Heidi Buhr, M.Sc.Med., Belinda J. Gabbe,
METHODS Ph.D., Doug W. Gould, Ph.D., Meg Harr-
We randomly assigned 750 adult patients in the ICU who were undergoing invasive old, Ph.D., Alisa M. Higgins, Ph.D., Sally
Hurford, P.G.Dip., Theodore J. Iwashyna,
mechanical ventilation to receive increased early mobilization (sedation minimiza- Ph.D., Ary Serpa Neto, Ph.D., Alistair D.
tion and daily physiotherapy) or usual care (the level of mobilization that was Nichol, Ph.D., Jeffrey J. Presneill, Ph.D.,
normally provided in each ICU). The primary outcome was the number of days that Stefan J. Schaller, M.D., Janani Sivasuthan,
M.P.H., Claire J. Tipping, Ph.D., Steven
the patients were alive and out of the hospital at 180 days after randomization. Webb, Ph.D., and Paul J. Young, M.B.,
Ch.B., Ph.D.) assume responsibility for the
RESULTS overall content and integrity of this article.
The median number of days that patients were alive and out of the hospital was 143
The affiliations of the members of the
(interquartile range, 21 to 161) in the early-mobilization group and 145 days (inter- writing committee are listed in the Ap-
quartile range, 51 to 164) in the usual-care group (absolute difference, −2.0 days; pendix. Dr. Hodgson can be contacted at
95% confidence interval [CI], −10 to 6; P = 0.62). The mean (±SD) daily duration of ­carol​.­hodgson@​­monash​.­edu or at the
Australian and New Zealand Intensive Care
active mobilization was 20.8±14.6 minutes and 8.8±9.0 minutes in the two groups, Research Centre, 553 St. Kilda Rd., Mel-
respectively (difference, 12.0 minutes per day; 95% CI, 10.4 to 13.6). A total of 77% bourne, VIC 3004, Australia.
of the patients in both groups were able to stand by a median interval of 3 days *A complete list of the TEAM trial inves-
and 5 days, respectively (difference, −2 days; 95% CI, −3.4 to −0.6). By day 180, death tigators and the ANZICS Clinical Trials
had occurred in 22.5% of the patients in the early-mobilization group and in 19.5% Group is provided in the Supplemen-
tary Appendix, available at NEJM.org.
of those in the usual-care group (odds ratio, 1.15; 95% CI, 0.81 to 1.65). Among
survivors, quality of life, activities of daily living, disability, cognitive function, and This article was published on October 26,
2022, at NEJM.org.
psychological function were similar in the two groups. Serious adverse events were
reported in 7 patients in the early-mobilization group and in 1 patient in the usual- N Engl J Med 2022;387:1747-58.
DOI: 10.1056/NEJMoa2209083
care group. Adverse events that were potentially due to mobilization (arrhythmias, Copyright © 2022 Massachusetts Medical Society.
altered blood pressure, and desaturation) were reported in 34 of 371 patients (9.2%)
in the early-mobilization group and in 15 of 370 patients (4.1%) in the usual-care CME
at NEJM.org
group (P = 0.005).
CONCLUSIONS
Among adults undergoing mechanical ventilation in the ICU, an increase in early
active mobilization did not result in a significantly greater number of days that pa-
tients were alive and out of the hospital than did the usual level of mobilization in
the ICU. The intervention was associated with increased adverse events. (Funded by the
National Health and Medical Research Council of Australia and the Health Research
Council of New Zealand; TEAM ClinicalTrials.gov number, NCT03133377.)

n engl j med 387;19 nejm.org november 10, 2022 1747

The New England Journal of Medicine
Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

A
pproximately 13 million to 20 mil- cal ventilation that was expected to continue be-
lion people worldwide receive treatment yond the calendar day after randomization. In the
in intensive care units (ICUs) annually.1 early-mobilization group, senior physiotherapists
ICU-acquired weakness, which is defined as clini- led the intervention and participated in interdisci-
A Quick Take is
cally detected weakness with no plausible explana- plinary discussions and reviews of a safety check-
available at tion except for critical illness,2 occurs in approxi- list. The trial was funded by the National Health
NEJM.org mately 40% of such patients3 and is associated and Medical Research Council of Australia and
with an increased risk of death, prolonged hos- the Health Research Council of New Zealand. The
pitalization, and impaired recovery.4-8 management committee designed the trial, which
Among patients in the ICU who have under- was endorsed by the Australian and New Zea-
gone mechanical ventilation for more than 48 land Intensive Care Society and the Irish Critical
hours, wasting of skeletal muscles occurs rapidly.9 Care Trials Group. The institutions that managed
Although immobilization may contribute to ICU- the trial and monitored data quality are listed in
acquired weakness, such weakness appears to be the Supplementary Appendix (available with the
part of the pathophysiology of critical illness and full text of this article at NEJM.org). An indepen-
is not just due to disuse.9 It is associated with dent data and safety monitoring committee over-
disruption of myofilament organization,10 damage saw the trial and reviewed a planned interim
to the sarcoplasmic reticulum, decreased electri- analysis after 400 patients had reached 28 days of
cal excitability, and mitochondrial dysfunction.11 follow-up. No commercial support was provided.
Although some data suggest that early mobiliza- The protocol (available at NEJM.org), which
tion of patients in the ICU may reduce the length was reported before the completion of enroll-
of hospital stay12 and improve function at the ment, was approved by the ethics committee at
time of hospital discharge,13,14 many barriers to each participating institution.20 Written informed
early mobilization exist.15-17 Moreover, early mo- consent for enrollment, or consent to continue
bilization may not be sufficient to prevent ICU- and to use data, was obtained from each patient
acquired weakness affecting patient-important or from a legal surrogate. In cases in which a pa-
outcomes and may be associated with risks.12,18 tient died before consent could be obtained, data
The Pain, Agitation–Sedation, Delirium, Immobil- were included if such inclusion was allowed by
ity, and Sleep Disruption in Adult Patients in the local regulations and approved by an ethics com-
ICU (PADIS) guidelines recommended mobiliza- mittee. The authors all contributed to the writing
tion of critically ill adults but do not offer advice of the article and the decision to submit the manu-
on the appropriate timing or regimen.19 script for publication. They also vouch for the
Accordingly, we conducted a clinical trial, called accuracy and completeness of the data and for
Treatment of Mechanically Ventilated Adults with the fidelity of the trial to the protocol.
Early Activity and Mobilization (TEAM), to test the
hypothesis that early active mobilization would Patients
increase the number of days that patients were Eligible patients were adults (≥18 years of age)
alive and out of the hospital at day 180 as com- who were expected to undergo mechanical ven-
pared with the usual level of mobilization in the tilation in the ICU beyond the calendar day after
ICU in adults who were undergoing mechanical randomization and whose condition was suffi-
ventilation. ciently stable to make mobilization potentially
possible. Key exclusion criteria were dependency
in any activity of daily living in the month before
Me thods
hospitalization, rest-in-bed orders, and proven or
Trial Design and Oversight suspected acute primary brain or spinal injury. A
In this international, multicenter, randomized, full list of the inclusion and exclusion criteria is
controlled trial, we evaluated the effects of early provided in the Supplementary Appendix. We de-
mobilization (sedation minimization and daily fined subgroups using baseline characteristics
physiotherapy) or usual care (mobilization level including prehospitalization disability level, age,
that was normally provided in each ICU) among illness severity, diagnosis, and frailty, as described
adults in the ICU who were undergoing mechani- in the Supplementary Appendix.

1748 n engl j med 387;19 nejm.org november 10, 2022

The New England Journal of Medicine

Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 Early Mobilization during Ventilation in the ICU

Randomization and Treatment Key secondary outcomes were mortality at

We randomly assigned patients in a 1:1 ratio to 180 days, the number of ventilator-free days and
receive early mobilization or usual care using a days out of the ICU from randomization to day
centralized Web-based interface. The trial statis- 28, and patient-reported outcome measures, in-
tician generated the assignment sequence using cluding quality of life and function in survivors
computer-generated random numbers stratified at day 180. We defined ventilator-free days as the
according to the trial center with variable block number of days of unassisted breathing during
sizes. the first 28 days after randomization; deaths by
We followed the guidelines of the Medical day 28 were assigned zero ventilator-free days.
Research Council for evaluating complex inter- ICU-free days were defined analogously.
ventions21 and the Template for Intervention We assessed health-related quality of life us-
Description and Replication.22 Our intervention, ing the EuroQol Group 5-Dimension Self-Report
which included minimization of sedation as re- Questionnaire (EQ-5D-5L), which evaluates mo-
quired, was hierarchical and began after ran- bility, personal care, usual activities, pain or dis-
domization with daily physiotherapy, which comfort, and anxiety or depression and catego-
could be provided in one or more sessions. The rizes each of these into five levels that range
sessions were individually tailored to achieve the from no problems to extreme problems,26 and
highest possible level of mobilization that was the EQ utility score, which ranges from −0.6 to
deemed to be safe for the patient at the initiation 1.0, with 1.0 indicating the best health state.27
of daily therapy. The highest level of mobiliza- The EQ Visual Analogue Scale provided a global
tion was provided for as long as possible before rating of patient-perceived health from 0 to 100,
a step-down to lower levels of activity if the pa- with higher scores indicating better health.28
tient became fatigued,23 as measured on the ICU Independent activities of daily living were mea-
Mobility Scale.24 This validated scale rates the sured with the Barthel Activities of Daily Living
level of mobilization from 0 to 10, with 0 indi- (ADL) Index, which scores 10 activities of daily
cating no mobilization and 10 indicating inde- living that include feeding, bathing, and dressing
pendent walking.25 Patients who were assigned from 0 (dependent) to 100 (independent),29 and
to the usual-care group received a level of mobi- the Lawton Instrumental Activities of Daily Liv-
lization that was normally provided at each site. ing Scale (IADL), which scores 8 independent
In both groups, concomitant care was guided by living skills that include shopping, laundry, and
treating clinicians. Details regarding the treat- housekeeping from 0 (dependent) to 8 (indepen-
ments and monitoring are provided in the Sup- dent).30,31 The 12-item World Health Organiza-
plementary Appendix. tion Disability Assessment Schedule (WHODAS
Patients received the trial treatment while they 2.0) measured generic function for mobility, self-
were in the ICU for up to 28 days after random- care, life activities, and participation.32 It incor-
ization. Blinding of mobilization in the ICU was porated scores from 0 (no difficulty) to 4 (ex-
not possible; however, trained staff who were treme difficulty) for each item on a scale of 0 to
unaware of trial-group assignments ascertained 48, with higher scores representing greater
patient-reported outcomes, and the statistical disability; this score is expressed as a percent-
analysis was performed in a blinded manner. age of maximum disability.32,33 Details regarding
additional outcomes, including cognitive and
Outcomes psychological function, are provided in the
The primary outcome was the number of days Supplementary Appendix.
that patients were alive and out of the hospital at Prespecified serious adverse events were falls
day 180. The time that patients were out of the to the floor, cardiac arrest, atrial fibrillation with
hospital was defined as the number days that they a ventricular response of more than 150 bpm,
were at home or in an accommodation that was other dangerous arrhythmias, oxygen-saturation
not a health care facility (i.e., a rehabilitation hos- level on pulse oximetry of less than 80% for more
pital or nursing home). Patients who had died by than 3 minutes, and unplanned extubation or re-
day 180 were defined as having zero days alive and moval of the intravascular line resulting in urgent
out of the hospital. replacement. Site investigators reported other ad-

n engl j med 387;19 nejm.org november 10, 2022 1749

The New England Journal of Medicine
Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

verse events that were potentially caused by ac- R e sult s

tive mobilization, as described in the Supplemen-
tary Appendix. Patients
From February 27, 2018, to November 19, 2021,
Statistical Analysis a total of 750 patients at 49 hospitals in 6 coun-
The statistical analysis plan was reported before tries underwent randomization to the early-mobi-
the completion of enrollment.20 The sample size lization group (372 patients) or usual-care group
was based on the standard deviation of the pri- (378 patients) (Fig. 1). Full consent was with-
mary outcome in the pilot study.14,34 We determined drawn by 1 patient in the early-mobilization group
that the enrollment of 750 patients would provide and 8 patients in the usual-care group, which left
90% power to detect a 7-day between-group differ- 371 in the early-mobilization group and 370 in
ence with a two-sided alpha of 0.05 after allowing the usual-care group. Three patients in the usu-
for 15% inflation to account for a nonparametric al-care group were lost to follow-up, and 5 pa-
distribution35 and 5% loss to follow-up. tients — 2 in the early-mobilization group and
The analysis of the primary outcome was per- 3 in the usual-care group — withdrew consent
formed in the intention-to-treat population, which for follow-up of the primary outcome but not for
was defined as all the patients who had been en- the use of other data. Accordingly, the primary
rolled except for those who had withdrawn consent outcome was available for 733 of 736 patients
for use of data. Between-group differences in the (99.6%) in the intention-to-treat population (369
primary outcome and continuous secondary out- in the early-mobilization group and 364 in the
comes were calculated with the use of median re- usual-care group). The trial groups had similar
gression with cluster-robust standard errors to ac- characteristics at baseline (Table 1 and Tables S1
count for site; results were reported as a difference through S5 in the Supplementary Appendix).
of medians with 95% confidence intervals.36 Fur- Table S6 describes the representativeness of the
ther analyses of the primary outcome included population.
evaluating the intervention effect across the range
of quantiles, ordinal logistic regression, and sensi- Activity and Mobilization
tivity to missingness. Prespecified subgroup analy- Mobilization milestones of active exercise, stand-
ses were conducted with the use of median regres- ing, and walking according to treatment group
sion with heterogeneity determined by fitting an are shown in Table 2 and Figure 2. The mean (±SD)
interaction between treatment assignment and daily duration of active mobilization per patient
subgroup with results reported in a forest plot. in the ICU was 20.8±14.6 minutes in the early-
Binomial secondary outcomes were compared with mobilization group and 8.8±9.0 minutes in the
the use of logistic regression to derive odds ratios usual-care group (absolute difference, 12 min-
with cluster-robust standard errors reported with utes per day; 95% confidence interval [CI], 10.4
95% confidence intervals, according to the statisti- to 13.6). Additional details regarding mobiliza-
cal analysis plan; however, odds ratios may over­ tion according to group are shown in Table S7
estimate the relative risk. and Figures S1, S2, and S3. Compliance with the
Analyses were conducted with the use of R intervention was high (Fig. S4). The major barri-
software, version 4.0.2,37 and SAS statistical ers to mobilization in the early-mobilization
software, version 9.4 (SAS Institute). Statistical group were protocol-compliant and included se-
significance for the primary outcome was de- dation, agitation, and physiological instability.
termined with the use of a two-sided hypothe-
sis test with an alpha of 0.05. Because there Primary Outcome
was no correction for multiple comparisons in At day 180, the median number of days that
secondary outcomes, P values have not been patients were alive and out of the hospital was
reported for these comparisons and the results 143 days (interquartile range, 21 to 161) in the
are considered to be exploratory. Multiple im- early-mobilization group and 145 days (interquar-
putation was performed for all outcomes with- tile range, 51 to 164) in the usual-care group
out complete data. Additional details regarding (absolute difference, −2 days; 95% CI, −10 to 6;
the analyses are provided in the Supplementary P = 0.62) (Table 3). Consistent with the results of
Appendix. the primary analysis were the findings in com-

1750 n engl j med 387;19 nejm.org november 10, 2022

The New England Journal of Medicine

Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 Early Mobilization during Ventilation in the ICU

10,828 Patients were assessed for eligibility

9736 Were excluded

954 Were dependent for activities
of daily living
304 Had cognitive impairment
4553 Had acute brain disease
482 Had spinal cord injury or
neuromuscular disease
1075 Had rest-in-bed orders
272 Had life expectancy <180 days
243 Were not recommended for
inclusion by physician
246 Had language barrier
485 Had been readmitted to ICU
440 Did not meet 72-hr cutoff for
randomization
682 Had other reason
342 Were eligible but did not undergo
randomization
70 Declined or had surrogate who
declined consent
272 Had other reason

750 Underwent randomization

372 Were assigned to early-mobilization group 378 Were assigned to usual-care group

14 Were excluded
3 Were excluded
8 Withdrew consent for
1 Withdrew consent for
all data
all data
3 Withdrew consent for
2 Withdrew consent for
follow-up at day 180
follow-up at day 180
3 Were lost to follow-up

369 Had primary outcome available at day 180 364 Had primary outcome available at day 180

Figure 1. Randomization and Outcomes.

In this trial involving patients who were undergoing invasive mechanical ventilation, early mobilization included se-
dation minimization and daily physiotherapy and usual care included the level of mobilization that was normally
provided in each intensive care unit (ICU).

plete-case analyses, sensitivity analyses for miss- and Fig. S6). The number of ventilator-free days
ingness, evaluation of the intervention effect esti- and ICU-free days at day 28 were similar in the
mate across quantiles and according to country, two groups.
and analyses of the primary outcome as an ordi- The scheduled 180-day follow-up occurred at
nal categorical variable (Tables S8 through S10 a median of 186 days (interquartile range, 182 to
and Fig. S5). 194) after randomization. Among the 579 survi-
vors, patient-reported outcomes were evaluated
Secondary Outcomes in 286 patients in the early-mobilization group
By day 180, deaths were reported in 83 of 369 and 293 in the usual-care group. The numbers of
patients (22.5%) in the early-mobilization group patients who completed each assessment and
and in 71 of 364 (19.5%) in the usual-care group findings regarding quality of life, activities of daily
(odds ratio, 1.15; 95% CI, 0.81 to 1.65) (Table 3 living, and disability were similar in the two

n engl j med 387;19 nejm.org november 10, 2022 1751

The New England Journal of Medicine
Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 1. Characteristics of the Patients at Baseline.*

Early Mobilization Usual Care

Characteristic (N = 371) (N = 370)
Age — yr 60.5±14.8 59.5±15.2
Female sex — no. (%) 128 (34.5) 146 (39.5)
Body-mass index† 29.9±7.9 30.4±7.8
Frailty and function
Median score on Clinical Frailty Scale (IQR)‡ 3 (2 to 4) 3 (2 to 4)
Median score on Functional Comorbidity Index (IQR)§ 2 (1 to 3) 2 (1 to 3)
Median score on WHODAS 2.0 (IQR)¶ 10.4 (2.1 to 25.0) 8.7 (2.1 to 22.7)
Highest score on the ICU Mobility Scale in wk before ICU admission‖ 9.9±0.6 9.8±0.7
Median interval from hospital admission to randomization (IQR) — hr 88.3 (50.5 to 137.0) 81.6 (48.2 to 147.0)
Median interval from ICU admission to randomization (IQR) — hr 60.1 (35 to 92.3) 61.3 (33.8 to 96.1)
ICU admission type — no. (%)
Planned ICU admission after elective surgery 68 (18.3) 58 (15.7)
Unplanned ICU admission 303 (81.7) 312 (84.3)
Median RASS score at randomization (IQR)** −3 (−4 to −2) −3 (−4 to −2)
Measurements and interventions at randomization††
Positive end-expiratory pressure — cm of water 8.9±3.0 8.8±3.1
Pao2:Fio2 226±79.1 230±85.2
Receipt of vasopressors by infusion — no. (%) 228 (61.5) 231 (62.4)
Receipt of renal-replacement therapy — no. (%) 82 (22.1) 79 (21.4)
APACHE II score‡‡ 18.2±6.8 18±6.9
Diagnosis subgroup — no. (%)§§
Sepsis¶¶ 246 (66.3) 245 (66.2)
Trauma 15 (4.0) 14 (3.8)
Covid-19 7 (1.9) 10 (2.7)

* Plus–minus values are means ±SD. Patients in the usual-care group received the level of mobilization that was normally
provided in each intensive care unit (ICU). Covid-19 denotes coronavirus disease 2019, and IQR interquartile range.
† The body-mass index is the weight in kilograms divided by the square of the height in meters.
‡ Scores on the Clinical Frailty Scale include 1 (very fit), 2 (well), 3 (managing well), 4 (vulnerable), 5 (mildly frail), 6 (mod-
erately frail), 7 (severely frail), 8 (very severely frail), and 9 (terminally ill). Patients were evaluated according to their
condition before the current admission, as confirmed by their next of kin or surrogate.
§ The Functional Comorbidity Index includes 18 diagnoses and scores from 0 to 18, with the score equal to the number
of specified coexisting illnesses present. Higher scores are associated with greater level of physical limitation.
¶ The 12-item World Health Organization Disability Assessment Schedule (WHODAS) 2.0 covers six domains of func-
tioning (with multiple questions for each domain), with scores ranging from 0 (no difficulty) to 4 (extreme difficulty)
and a total score ranging from 0 to 48, with higher scores representing greater disability. The total score is divided
by 48 and multiplied by 100 to convert it to a percentage of maximum disability. The WHODAS 2.0 score was avail-
able for 322 patients in the early-mobilization group and 310 patients in the usual-care group. At randomization, the
WHODAS 2.0 was completed by the patient’s next of kin or surrogate.
‖ Scores on the ICU Mobility Scale include 0 (lying in bed), 1 (sitting and exercising in bed), 2 (passive movement from
bed to chair, no standing), 3 (sitting on edge of bed), 4 (standing), 5 (transferring from bed to chair), 6 (marching in
place at bedside), 7 (walking with assistance of 2 or more people), 8 (walking with assistance of 1 person), 9 (walk-
ing independently with gait aid), and 10 (walking independently without gait aid). The score was obtained from the
patient’s next of kin or surrogate at the time of randomization.
** Scores on the Richmond Agitation Sedation Scale (RASS) range from −5 (unarousable) to 4 (combative). A score of
−4 to −2 indicates deep to light sedation. This score was available for 358 patients in the early-mobilization group and
357 patients in the usual-care group.
†† Data regarding positive end-expiratory pressure were available for all the patients in the early-mobilization group and
369 patients in the usual-care group. Data regarding the ratio of arterial partial pressure of oxygen (Pao2) to the frac-
tion of inspired oxygen (Fio2) were available for 369 patients and 368 patients, respectively.
‡‡ Scores on the Acute Physiology and Chronic Health Evaluation (APACHE) II range from 0 to 71, with higher scores
indicating more severe disease and a higher risk of death.
§§ Subgroups of this category were the only prespecified diagnoses that were evaluated and were not mutually exclusive.
Data regarding the ICU admission source and diagnosis are provided in Table S4.
¶¶ Sepsis was defined as suspected or confirmed infection plus a score on the Sequential Organ Failure
Assessment (SOFA) of 2 or more if there was no known preexisting organ dysfunction or an increase from baseline in
the SOFA score of more than 2 points if there was preexisting organ dysfunction.

1752 n engl j med 387;19 nejm.org november 10, 2022

The New England Journal of Medicine

Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 Early Mobilization during Ventilation in the ICU

Table 2. Mobilization in the ICU.*

Early Usual Between-Group

Mobilization Care Difference
Characteristic (N = 371) (N = 370) (95% CI)†
Patients who were assessed by a physiotherapist on day of 320/370 (86.5) 265/363 (73.0) 13.5 (6.7 to 20.3)
randomization — no./total no. (%)
No. of days per patient when physiotherapy assessment oc- 0.94±0.11 0.81±0.24 0.14 (0.12 to 0.16)
curred
No. of minutes of active mobilization per day 20.8±14.6 8.8±9.0 12.0 (10.4 to 13.6)
Mobilization milestones‡
IMS 3 or higher
Patients — no. (%) 331 (89.2) 330 (89.2) 0 (−4.3 to 4.3)
Median no. of days since randomization (IQR) 3 (1 to 6) 4 (2 to 7) −1 (−2.2 to −0.2)
IMS 4 or higher
Patients — no. (%) 287 (77.4) 286 (77.3) 0.1 (−6.0 to 6.1)
Median no. of days since randomization (IQR) 3 (2 to 7) 5 (3 to 8) −2 (−3.4 to −0.6)
IMS 7 or higher
Patients — no. (%) 176 (47.4) 150 (40.5) 6.9 (−0.2 to 14.0)
Median no. of days since randomization (IQR) 5 (3 to 8) 7 (4 to 13) −2 (−3.4 to −0.7)
Median peak IMS (IQR) 6 (4 to 8) 6 (4 to 8) 0 (−1 to 1)

* Plus–minus values are means ±SD.

† Between-group differences were calculated after adjustment for trial site.
‡ A score of 3 on the ICU mobility scale (IMS 3) (sitting on the edge of the bed) could involve assistance of a staff
member but required that the patient was actively sitting with some trunk control. IMS 4 (standing) required weight
bearing through the feet with or without assistance or the use of a standing lifter or tilt-table device. IMS 5 (transfer-
ring from bed to chair) required active transfer of weight from one leg to another to move to the chair; if standing was
accomplished with the assistance of a medical device, stepping into the chair was required. IMS 6 (marching in place)
required that the patient lift alternate feet at least four times (twice on each foot) with or without assistance. IMS 7
(walking with assistance of ≥2 people), IMS 8 (walking with the assistance of 1 person), IMS 9 (walking independently
with a gait aid), and IMS 10 (walking independently without a gait aid) required walking at least 5 m away from the bed
or chair.

groups (Tables S11 and S12). Additional second- Safety Outcomes

ary outcomes, including 28-day mortality and Adverse events that were potentially due to mo-
cognitive and psychological function, were also bilization were reported in 34 of 371 patients
similar (Table S13). (9.2%) in the early-mobilization group and in 15
of 370 (4.1%) in the usual-care group (P = 0.005);
Process of Care Measures and Subgroup cardiac arrhythmia, altered blood pressure, and
Analyses oxygen desaturation were the most commonly
Process-of-care measures — which included the reported events (Table 3). A total of 8 serious
use of tracheostomy, neuromuscular blockers, glu- adverse events were reported, of which 7 occurred
cocorticoids, new renal-replacement therapy, re- in the early-mobilization group (5 arrhythmias, a
intubation, and vasopressor-free days — were desaturation episode, and a cerebrovascular ac-
similar in the two groups (Table S14). Daily se- cident), and 1 occurred in the usual-care group
dation scores are shown in Figure S7, and the (a desaturation episode) (Table S15). All serious
proportion of patients in the ICU with delirium adverse events required medical intervention. All
each day is shown in Figure S8. Daily ventilatory events resolved except for the cerebrovascular
support is shown in Figure S9. There was no accident, which resulted in persistent unilateral
evidence of heterogeneity in the effect of early weakness. No instances of falling to the floor,
mobilization on the primary outcome among cardiac arrest, unplanned extubation, or intra-
patients in any of the prespecified subgroups vascular line removal resulting in urgent replace-
(Fig. S10). ment were reported.

n engl j med 387;19 nejm.org november 10, 2022 1753

The New England Journal of Medicine
Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

ICU Mobility Scale: 0 (nothing or 1–2 (in-bed or 3–4 (active sitting 5–6 (transfer or 7–8 (assisted 9–10 (independent
passive) in-chair exercises) or standing) marching in place) walking) walking)

A Early Mobilization
100

90

80
Percentage of Patients

70

60

50

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Day
No. of Patients 365 371 367 342 306 281 251 221 202 177 165 147 131 117 109 104 98 84 78 72 64 59 49 48 47 43 40 38

B Usual Care
100

90

80
Percentage of Patients

70

60

50

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Day
No. of Patients 363 364 362 347 311 281 254 236 204 184 166 146 134 125 112 104 96 87 81 78 69 61 59 56 52 47 43 41

Figure 2. Mobilization in the Intensive Care Unit, According to Treatment Group.

Discussion cantly increased the number of days that pa-

tients were alive and out of the hospital at 180
In this international, randomized, controlled trial days.14 However, in this meta-analysis, both the
involving adults who were undergoing mechani- intensity and duration of mobilization in the
cal ventilation, the numbers of days that patients control groups varied greatly, a factor that made
were alive and out of the hospital at 180 days it difficult to draw comparisons across the trials.
were similar in the early-mobilization group and Our trial avoided some of the methodologic short-
the usual-care group. Adverse events and serious comings of studies that were included in this meta-
adverse events were reported more commonly in analysis, such as small sample sizes,34,38 single-
the early-mobilization group. center designs,39 and use of historical controls.39
Our findings are at variance with a meta- Some randomized, controlled trials of early
analysis showing that active mobilization in the mobilization that have shown potential benefits
ICU, particularly when delivered early, signifi- have focused on outcomes other than days alive

1754 n engl j med 387;19 nejm.org november 10, 2022

The New England Journal of Medicine

Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 Early Mobilization during Ventilation in the ICU

Table 3. Primary Outcome, Key Secondary Outcomes, and Adverse Events.*

Early Mobilization Usual Care Difference or Odds

Outcome (N = 371) (N = 370) Ratio (95% CI)† P Value
Primary outcome
Days alive and out of hospital at day 180‡
Median no. (IQR) 143 (21 to 161) 145 (51 to 164) −2.0 (−10 to 6) 0.62
Key secondary outcomes
Death at day 180
Patients — no. (%) 83/369 (22.5) 71/364 (19.5) 1.15 (0.81–1.65)§
Median no. of days since randomization (IQR) 17 (9 to 41) 19 (12 to 50) −2.0 (−12.0 to 8.0)
Median no. of ventilator-free days at day 28 (IQR) 21 (8 to 25) 21 (11 to 25) 0.0 (−1.4 to 1.4)
Median no. of ICU-free days at day 28 (IQR) 16 (0 to 21) 17 (3 to 22) −1.0 (−3.1 to 1.1)
Functional outcomes in survivors at day 180¶
Score on EQ-5D-5L utility score‖ 0.7±0.3 0.7±0.3 0.0 (−0.0 to 0.1)
Score on EQ Visual Analogue Scale** 70.2±19.7 69.0±20.1 2.0 (−5.7 to 9.7)
Median score on Barthel Index of ADL (IQR)†† 100 (100 to 100) 100 (95 to 100) 0
Median score on IADL (IQR)‡‡ 8.0 (7.0 to 8.0) 8.0 (6.0 to 8.0) 0.2 (−0.9 to 1.3)
Median score on WHODAS 2.0 (IQR)§§ 12.5 (2.1 to 33.3) 14.6 (4.2 to 38.9) −1.8 (−6.9 to 3.4)
Adverse events — no. (%)¶¶
Patients with ≥1 adverse event potentially due to mobilization 34 (9.2) 15 (4.1) 2.55 (1.33–4.89)§ 0.005
— no. (%)
Adverse events per patient — no. (%) 0.02
0 337 (90.8) 355 (95.9)
1 19 (5.1) 11 (3.0)
2 4 (1.1) 2 (0.5)
≥3 11 (3.0) 2 (0.5)
Type of adverse events — no. (%)‖‖
Altered blood pressure 13 (3.5) 8 (2.2) 0.27
Cardiac arrhythmia 13 (3.5) 4 (1.1) 0.03
Oxygen desaturation 8 (2.2) 1 (0.3) 0.02
Pain or agitation 4 (1.1) 1 (0.3) 0.37
Removal of invasive line 2 (0.5) 2 (0.5) 1.00
Gastrointestinal 2 (0.5) 1 (0.3) 1.00
Tachypnea 3 (0.8) 0 0.25
Altered neurologic state 1 (0.3) 1 (0.3) 1.00
Other 4 (1.1) 0 0.12

* Plus–minus values are means ±SD.

† Values are between-group differences unless otherwise indicated. All differences in medians were calculated with the use of quantile regres-
sion after adjustment for trial site. Multiple imputation for missingness was used for primary and key secondary outcomes.
‡ Data for the primary and secondary outcomes were available for 733 patients (369 in the early-mobilization group and 364 in the usual-care group).
§ This value is an odds ratio that was calculated after adjustment for the trial site as a random effect with the use of multiple imputation to
account for missingness.
¶ Of 579 survivors, 286 in the early-mobilization group and 293 in the usual-care group were contacted to complete functional outcome assessments.
‖ The EuroQol Group 5-Dimension Self-Report Questionnaire (EQ-5D-5L) includes of five dimensions: mobility, self-care, usual activities,
pain or discomfort, and anxiety or depression. Scores on the utility scale range from −0.6 to 1, with the maximum score indicating the best
health state. The utility score was missing for 38 patients in the early-mobilization group and 38 in the usual-care group.
** The EQ Visual Analogue Scale provides a single global rating of self-perceived health and is scored on a scale of 0 to 100, with higher scores
indicating better health. Data on this scale were missing for 40 patients in the early-mobilization group and 39 in the usual-care group.
†† The Barthel Index of Activities of Daily Living (ADL) measures functional disability in 10 ADLs by quantifying patient performance. Increments
of 5 points are used in scoring, with a maximal score of 100 indicating full independence in physical functioning and a lowest score of 0 indi-
cating a completely bed-bound state. The ADL was missing for 39 patients in the early-mobilization group and 43 in the usual-care group.
‡‡ The Lawton Instrumental Activities of Daily Living (IADL) Scale is an assessment of independent living skills across eight domains of
function. A summary score ranges between from 0 and 8 with higher scores indicating greater levels of independence. The IADL score was
missing for 37 patients in the early-mobilization group and 43 in the usual-care group.
§§ The WHODAS 2.0 score was missing for 52 patients in the early-mobilization group and 52 in the usual-care group.
¶¶ Adverse events include events that were reported as probably, possibly, or definitely related to mobilization. These do not include serious
adverse events, which are reported separately in Table S15.
‖‖ In cases in which a patient had a particular adverse event on more than one occasion, each event was only counted once.

n engl j med 387;19 nejm.org november 10, 2022 1755

The New England Journal of Medicine
Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

and out of the hospital. In a trial that involved mobilization in the intervention group, but we did
patients being treated in a medical ICU, early not record whether this occurred. Despite sepa-
mobilization with an interruption of sedation ration between groups in timing and duration of
increased the level of independent function at mobilization, factors including sedation, agita-
the time of hospital discharge.13 In that trial, tion, and physiological instability often precluded
mobilization milestones in the usual-care group mobilization. This meant that some patients were
were met at a similar time as were those in our not actively mobilized in the ICU. Patients who
trial. However, in contrast to the previous trial, were not mobilized reduced the statistical power
which began mobilization sessions with low lev- to detect a difference between groups. Our proto-
els of activity, we sought to begin mobilizing col did not stipulate details regarding rehabilita-
patients at the highest level of activity possible. tion beyond the ICU or call for an assessment of
In one study involving patients in a surgical ICU, function at hospital discharge. The observation
early mobilization reduced the length of stay in of a greater frequency of adverse events with early
the ICU and increased functional mobility at hos- mobilization than with usual care may have been
pital discharge.12 However, as in our trial, patients subject to surveillance bias because the treatment
in the early-mobilization group in that trial had an assignments were unblinded. In contrast, for pa-
increased risk of adverse events. Moreover, the tient-reported outcomes at day 180, we used cen-
interpretation of that study in surgical patients tralized assessors who were unaware of trial-
was complicated by markedly higher in-hospital group assignments to avoid bias. Because these
mortality in the early-mobilization group. outcomes could be compared only among survi-
Our findings are broadly consistent with the vors, they do not represent randomized com-
results of three randomized, controlled trials that parisons. Some data related to patient-reported
were conducted during the past 6 years.40-42 In two outcomes at day 180 were missing. These data
of these trials, investigators compared intensive may not have been missing at random because
physiotherapy with usual care among patients in patients with better (or worse) outcomes might
the ICU and reported no between-group difference have been harder to contact or less likely to com-
among survivors with respect to physical function plete interviews. For our primary outcome, miss-
at 1 month and at 6 months, respectively.40,41 The ing data were rare and our findings were consis-
third trial compared standardized rehabilitation tent in a range of sensitivity analyses. Odds ratio
therapy with usual ICU care in adults with acute may overestimate the relative risk and be mis-
respiratory failure and showed no difference be- interpreted.
tween groups in the length of hospital stay.42 Mobilizing critically ill patients early requires
Our trial has some limitations. Patients in our clinical expertise, time, and resources. Although
usual-care group received a higher level of mobi- we used a safety checklist,23 conducted interdis-
lization than those in some cohort studies43,44 and ciplinary discussion with the medical team, and
in the control groups of some previous trials.12,41 required that senior physiotherapists direct early
However, the mobilization levels that were achieved active mobilization, our trial suggests greater
in our usual-care group were consistent with those safety with usual care than with the additional
outlined in international guidelines,19,45 were simi- early mobilization that was provided in our trial.
lar to those in a recent multicenter cohort study,46 Thus, for adults undergoing mechanical ventila-
and were similar to those in the control group tion in the ICU, increased early active mobilization
of a previous clinical trial showing benefits from did not affect the number of days that they were
early mobilization.13 Although practitioners in alive and out of the hospital as compared with the
our trial provided treatments according to the usual level of mobilization received in the ICU.
protocol, changes in practice that could have af-
fected treatment in the usual-care group may have
Supported by the National Health and Medical Research
occurred in specific sites or countries. Our proto- Council of Australia and the Health Research Council of New
col stipulated that whenever it was feasible to do Zealand.
so, patients in the usual-care group should re- Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
ceive treatment from physiotherapy staff mem- A data sharing statement provided by the authors is available
bers who were not involved in delivering early with the full text of this article at NEJM.org.

1756 n engl j med 387;19 nejm.org november 10, 2022

The New England Journal of Medicine

Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 Early Mobilization during Ventilation in the ICU

Appendix
The affiliations of the members of the writing committee are as follows: the Australian and New Zealand Intensive Care Research Cen-
tre, Department of Epidemiology and Preventive Medicine (C.L.H., M.B., R.B., T.B., A.M.H., A.S.N., A.D.N., J.J.P., J.S., S.W., P.J.Y.),
School of Public Health and Preventive Medicine (B.J.G.), Monash University, the Data Analytics Research and Evaluation Centre, Uni-
versity of Melbourne and Austin Hospital (R.B., A.S.N.), the Department of Critical Care (C.L.H., R.B., A.S.N., J.J.P., P.J.Y.) and the
School of Medicine (J.J.P.), University of Melbourne, the Department of Intensive Care (A.D.N.) and the Intensive Care Unit and Phys-
iotherapy Department (C.L.H., C.J.T.), Alfred Hospital, and the Department of Intensive Care, Royal Melbourne Hospital (R.B., J.J.P.),
Melbourne, VIC, the Critical Care Division, the George Institute for Global Health (C.L.H., A.M.H.), and Intensive Care Services, Royal
Prince Alfred Hospital (H.B.), Sydney, the Curtin School of Allied Health, Curtin University, Bentley, WA (M.H.), and the Department
of Physiotherapy, Royal Perth Hospital (M.H.), and the Intensive Care Unit, St. John of God Subiaco Hospital (S.W.), Perth, WA — all
in Australia; the Intensive Care Unit, Wellington Hospital (P.J.Y.), and the Medical Research Institute of New Zealand (S.H., P.J.Y.) —
both in Wellington, New Zealand; the Department of Anesthesiology and Intensive Care, School of Medicine, Technical University of
Munich, School of Medicine, Klinikum Rechts der Isar, Munich, and the Department of Anesthesiology and Operative Intensive Care
Medicine, Charité–Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin
— both in Germany (S.J.S); the Department of Internal Medicine Division of Pulmonary and Critical Care, University of Michigan, Ann
Arbor (T.J.I.); the Department of Medicine Division of Pulmonary and Critical Care, Johns Hopkins University, Baltimore (T.J.I.); the
Clinical Trials Unit, Intensive Care National Audit and Research Centre, London (D.W.G.); the Department of Critical Care Medicine,
Hospital Israelita Albert Einstein, Sao Paulo (A.S.N.); and University College Dublin–Clinical Research Centre at St. Vincent’s Univer-
sity Hospital, Dublin (K.B., A.D.N.).

References
1. Adhikari NK, Fowler RA, Bhagwanjee sive care unit-acquired weakness. Eur the phase 3 randomised controlled Treat-
S, Rubenfeld GD. Critical care and the Respir J 2012;​39:​1000-11. ment of Invasively Ventilated Adults with
global burden of critical illness in adults. 12. Schaller SJ, Anstey M, Blobner M, et al. Early Activity and Mobilisation (TEAM
Lancet 2010;​376:​1339-46. Early, goal-directed mobilisation in the III) trial. Crit Care Resusc 2021;​23:​262-72
2. Kress JP, Hall JB. ICU-acquired weak- surgical intensive care unit: a randomised (https://ccr​.­cicm​.­org​.­au/​­journal​-­editions/​
ness and recovery from critical illness. controlled trial. Lancet 2016;​388:​1377-88. ­articles/​­1439).
N Engl J Med 2014;​370:​1626-35. 13. Schweickert WD, Pohlman MC, Pohl- 21. Craig P, Dieppe P, Macintyre S, Michie
3. Appleton RT, Kinsella J, Quasim T. man AS, et al. Early physical and occupa- S, Nazareth I, Petticrew M. Developing
The incidence of intensive care unit-ac- tional therapy in mechanically ventilated, and evaluating complex interventions: the
quired weakness syndromes: a systematic critically ill patients: a randomised con- new Medical Research Council guidance.
review. J Intensive Care Soc 2015;​16:​126-36. trolled trial. Lancet 2009;​373:​1874-82. BMJ 2008;​337:​a1655.
4. Fan E, Dowdy DW, Colantuoni E, et al. 14. Tipping CJ, Harrold M, Holland A, 22. Hoffmann TC, Glasziou PP, Boutron
Physical complications in acute lung in- Romero L, Nisbet T, Hodgson CL. The ef- I, et al. Better reporting of interventions:
jury survivors: a two-year longitudinal fects of active mobilisation and rehabili- template for intervention description and
prospective study. Crit Care Med 2014;​42:​ tation in ICU on mortality and function: a replication (TIDieR) checklist and guide.
849-59. systematic review. Intensive Care Med BMJ 2014;​348:​g1687.
5. Hermans G, Van Mechelen H, Clerckx 2017;​43:​171-83. 23. Hodgson CL, Stiller K, Needham DM,
B, et al. Acute outcomes and 1-year mor- 15. Cuthbertson BH, Goddard SL, Loren- et al. Expert consensus and recommenda-
tality of intensive care unit-acquired catto F, et al. Barriers and facilitators to tions on safety criteria for active mobili-
weakness: a cohort study and propensity- early rehabilitation in the ICU: a theory zation of mechanically ventilated criti-
matched analysis. Am J Respir Crit Care driven Delphi study. Crit Care Med 2020;​ cally ill adults. Crit Care 2014;​18:​658.
Med 2014;​190:​410-20. 48(12):​e1171-e1178. 24. Hodgson C, Needham D, Haines K,
6. De Jonghe B, Sharshar T, Lefaucheur 16. Dubb R, Nydahl P, Hermes C, et al. et al. Feasibility and inter-rater reliability
J-P, et al. Paresis acquired in the intensive Barriers and strategies for early mobiliza- of the ICU Mobility Scale. Heart Lung
care unit: a prospective multicenter study. tion of patients in intensive care units. 2014;​43:​19-24.
JAMA 2002;​288:​2859-67. Ann Am Thorac Soc 2016;​13:​724-30. 25. Tipping CJ, Bailey MJ, Bellomo R, et al.
7. Ali NA, O’Brien JM Jr, Hoffmann SP, 17. Hodgson CL, Capell E, Tipping CJ. The ICU Mobility Scale has construct and
et al. Acquired weakness, handgrip Early mobilization of patients in intensive predictive validity and is responsive:
strength, and mortality in critically ill pa- care: organization, communication and a multicenter observational study. Ann Am
tients. Am J Respir Crit Care Med 2008;​ safety factors that influence translation Thorac Soc 2016;​13:​887-93.
178:​261-8. into clinical practice. Crit Care 2018;​22:​ 26. Herdman M, Gudex C, Lloyd A, et al.
8. Van Aerde N, Meersseman P, Debav- 77. Development and preliminary testing of
eye Y, et al. Five-year impact of ICU-acquired 18. Zhang L, Hu W, Cai Z, et al. Early mo- the new five-level version of EQ-5D (EQ-
neuromuscular complications: a prospec- bilization of critically ill patients in the 5D-5L). Qual Life Res 2011;​20:​1727-36.
tive, observational study. Intensive Care intensive care unit: a systematic review 27. Gerth AMJ, Hatch RA, Young JD, Wat-
Med 2020;​46:​1184-93. and meta-analysis. PLoS One 2019;​14(10):​ kinson PJ. Changes in health-related
9. Puthucheary ZA, Rawal J, McPhail M, e0223185. quality of life after discharge from an
et al. Acute skeletal muscle wasting in 19. Devlin JW, Skrobik Y, Gélinas C, et al. intensive care unit: a systematic review.
critical illness. JAMA 2013;​310:​1591-600. Clinical practice guidelines for the pre- Anaesthesia 2019;​74:​100-8.
10. Lad H, Saumur TM, Herridge MS, et al. vention and management of pain, agita- 28. Rabin R, de Charro F. EQ-5D: a mea-
Intensive care unit-acquired weakness: tion/sedation, delirium, immobility, and sure of health status from the EuroQol
not just another muscle atrophying condi- sleep disruption in adult patients in the Group. Ann Med 2001;​33:​337-43.
tion. Int J Mol Sci 2020;​21:​7840. ICU. Crit Care Med 2018;​46(9):​e825-e873. 29. da Silveira LTY, da Silva JM, Soler JMP,
11. Bloch S, Polkey MI, Griffiths M, 20. Presneill JJ, Bellomo R, Brickell K, et al. Sun CYL, Tanaka C, Fu C. Assessing func-
Kemp P. Molecular mechanisms of inten- Protocol and statistical analysis plan for tional status after intensive care unit stay:

n engl j med 387;19 nejm.org november 10, 2022 1757

The New England Journal of Medicine
Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.
 Early Mobilization during Ventilation in the ICU

the Barthel Index and the Katz Index. Int tical methods based on ranks. New York:​ (EPICC): a multicentre, parallel-group,
J Qual Health Care 2018;​30:​265-70. Springer-Verlag;​ 2006. randomised controlled trial. Thorax 2018;​
30. Graf C. The Lawton instrumental ac- 36. Koenker R, Portnoy S, Ng PT, et al. 73:​213-21.
tivities of daily living scale. Am J Nurs Quantreg: Quantile regression. R package, 42. Morris PE, Berry MJ, Files DC, et al.
2008;​108:​52-62. version 5.94. July 20, 2022 (https://CRAN​.­R​ Standardized rehabilitation and hospital
31. Hopkins RO, Suchyta MR, Kamdar -­project​.­org/​­package=​­quantreg). length of stay among patients with acute
BB, Darowski E, Jackson JC, Needham 37. R Core Team. The R Project for statis- respiratory failure: a randomized clinical
DM. Instrumental activities of daily liv- tical computing. Vienna:​R Foundation for trial. JAMA 2016;​315:​2694-702.
ing after critical illness: a systematic re- Statistical Computing, 2019 (https://www​ 43. Harrold ME, Salisbury LG, Webb SA,
view. Ann Am Thorac Soc 2017;​14:​1332- .­R​-­project​.­org/​­). Allison GT. Early mobilisation in inten-
43. 38. Wolfe KS, Wendlandt BN, Patel S, et al. sive care units in Australia and Scotland:
32. Higgins AM, Serpa Neto A, Bailey M, Long-term survival and health care utili- a prospective, observational cohort study
et al. The psychometric properties and zation of mechanically ventilated patients examining mobilisation practises and
the minimal clinically important differ- in a randomized controlled trial of early barriers. Crit Care 2015;​19:​336.
ence for disability assessment using the mobilization. Am J Respir Crit Care Med 44. The TEAM Study Investigators. Early
WHODAS 2.0 in critically ill patients. 2013;​187:​A5235. abstract. mobilization and recovery in mechani-
Crit Care Resusc 2021;​23:​103-12 (https:// 39. Hanekom SD, Louw Q, Coetzee A. cally ventilated patients in the ICU: a bi-
ccr​.­cicm​.­org​.­au/​­journal​-­editions/​­a rticles/​ The way in which a physiotherapy service national, multi-centre, prospective cohort
­1411). is structured can improve patient out- study. Crit Care 2015;​19:​81.
33. Higgins AM, Neto AS, Bailey M, et al. come from a surgical intensive care: a 45. Lang JK, Paykel MS, Haines KJ, Hodg-
Predictors of death and new disability af- controlled clinical trial. Crit Care 2012;​ son CL. Clinical practice guidelines for
ter critical illness: a multicentre prospec- 16:​R230. early mobilization in the ICU: a system-
tive cohort study. Intensive Care Med 40. Moss M, Nordon-Craft A, Malone D, atic review. Crit Care Med 2020;​ 48(11):​
2021;​47:​772-81. et al. A randomized trial of an intensive e1121-e1128.
34. Hodgson CL, Bailey M, Bellomo R, physical therapy program for patients 46. Raurell-Torredà M, Arias-Rivera S,
et al. A binational multicenter pilot feasi- with acute respiratory failure. Am J Respir Martí JD, et al. Variables associated with
bility randomized controlled trial of early Crit Care Med 2016;​193:​1101-10. mobility levels in critically ill patients:
goal-directed mobilization in the ICU. 41. Wright SE, Thomas K, Watson G, et al. a cohort study. Nurs Crit Care 2022;​27:​
Crit Care Med 2016;​44:​1145-52. Intensive versus standard physical reha- 546-57.
35. Lehmann EL. Nonparametrics: statis- bilitation therapy in the critically ill Copyright © 2022 Massachusetts Medical Society.

1758 n engl j med 387;19 nejm.org november 10, 2022

The New England Journal of Medicine

Downloaded from nejm.org on May 30, 2023. For personal use only. No other uses without permission.
Copyright © 2022 Massachusetts Medical Society. All rights reserved.