Journal of Intensive and Critical Care Open Access

Keywords

Critical care; Mobilization; Propensity

Introduction

Patients admitted to Intensive care units (ICU) usually undergo unavoidable periods of immobilization pertaining to the nature of interventions required as part of their management [1], interventions such as mechanical ventilation (MV), sedation, along with the insertion of multiple devices to provide fluids, nutrition, vasopressors and medication or for monitoring purposes, such as central lines, endotracheal tubes, nasogastric tubes, Foley’s catheters and others [2]. Immobilization of critically ill patients, although essential for the management, is not without sequel, immobilization decreases muscle strength by 20% within the first week and further 20% for every subsequent week of immobilization [3], it has been associated with ICU acquired weakness [4], in addition to atelectasis, bone demineralization, vasomotor instability, pressure ulcers, contractures and venous thromboembolism [5]. All these impacts result essentially in difficulty of weaning from MV, prolonged ICU and hospital length of stay (LOS) which is subsequently translated in increased daily costs and resources utilization [6].

Routine physiotherapy (PT) in ICU although intended to counter such sequel, has not been determined to benefit patients [1], furthermore, study show that only 27% of ICU patients receive such service, only 6% of which are mechanically ventilated [7]. Early mobilization of critically ill and mechanically ventilated patients was first described in 2007 by Bailey et al. [8], in their study they concluded the intervention’s safety and effectiveness. Since then early mobilization of critically ill patients was studied extensively, a recent meta-analysis of 17 studies concluded its beneficial effects [9].

Aim of the Study

To report and evaluate the results of an early mobilization program for critically ill patients, as a quality improvement (QI) project.

Study Design

The early mobilization of critically ill patients program is a QI project applied in the ICU at King Saud Medical City, the largest ministry of health hospital in Saudi Arabia, with 127 operational beds, divided into surgical, medical, trauma and isolation units. It was approved by the total quality management (TQM) in our institution.

The project started by developing a protocol for the program, which is mostly based on the framework of the mobilization program devised by Priyakshi et al. [1]. In our protocol the core team consisted of critical care physician, critical care nurse, respiratory therapist, physiotherapist and patient safety and quality coordinator.

Patients in the ICU were screened daily for inclusion criteria, which were: The patient was independently able to mobilize prior to the current hospital admission (this includes patients who use a walking stick or gait aid to mobilize, but not patients who need assistance from another person or a machine such as a wheelchair), the team is able to communicate with the patient in a language that he/she understands, age 14 years or more, hemodynamically stable (mean arterial pressure [MAP]≥60 mmHg without vasopressors or inotropes), fully conscious, with no use of sedatives, or muscle relaxants, maintains oxygen saturation of 95% or more with FiO2 not more than 50% whether they were mechanically ventilated or not.

Whereas the exclusion criteria were: Complete bed rest order, unstable fractures, palliative care/DNR order, active bleeding, evidence of increased intracranial pressure (by monitoring, radiological, or clinical criteria), surgically open abdomen, mechanically ventilated with PEEP more than 12, or controlled mode of ventilation of any setting, respiratory Rate≥40 /min, acidosis with PH ≤ 7.25, tachycardia ≥ 130 beats/min, ongoing renal replacement therapy and ventilator dysynchrony.

Other than the exclusion criteria, there were certain safety considerations that had to be discussed among the core team members for each patient individually, in order to decide whether to exclude the patient permanently from the program or to wait for improvement (if applicable), these considerations were: Body Mass Index>40, hemoglobin<7 g/dl, recent tracheostomy within 24 h, history of cardiac arrest/cardiopulmonary resuscitation, venous thromboembolism, multiple lines and drains and response of the patient to the previous session of mobilization if already enrolled.

The program included six levels of activities (Table 1) and an algorithm to move the patient up and down between these levels according to his/her response. The protocol also details the maximum duration of each session (20 min), when to interrupt a session, required instruments and material, infection control considerations and key performance indicators (KPIs) to be measured (full protocol in supplementary file). The program was officially launched on January 1st, 2017, since the program was an approved QI; no consent was required to enroll patients.

Table 1: Mobilization program levels of activities.

Apart from enrollment in the early mobilization program all patients were managed similarly according to our ICU protocols specific for the patient’s conditions, the ICU protocols are adapted from the best available knowledge and are based on evidencebased medicine, for example the brain trauma foundation guidelines [10] for traumatic brain injury and surviving sepsis campaign guidelines [11] for septic shock.

Outcomes

The primary outcome in this report is the effect of program enrollment on the LOS in ICU, through propensity score matching of enrolled patients as compared to patients not enrolled. Secondary outcomes include: average LOS, average duration of weaning trials and ICU mortality.

Statistical Method

For the primary outcome of propensity score matching a logistic regression model was developed using enrollment in the program as the dependent variable, independent variables were age, gender, APACHE 4 score, source of ICU admission (ED or ward) and binary diagnostic general category as medical or surgical, using enter method and Hosmer Lemeshow test to evaluate goodness of fit (well fitted if p>0.05), the model was used to identify variables which will be used to generate propensity scores and those are the variables with p<0.2 in the logistic regression model. Propensity scores were used to match enrolled patients in the program with nearest neighbors from the control group with a maximum difference of 0.005 between matched scores, in a 1:1 ratio. Propensity score matching aims at adjusting any imbalances between the compared groups that may arise from the absence of randomization in the treatment assignment, in addition to reduction of bias and variance in the estimation of treatment effect. Results were reported as average effect of treatment. This statistical method is based on that described by Brookhart et al. [12] and has been utilized by other researchers to control for confounding factors [13].

Baseline characteristics of the study population and the secondary outcomes were summarized as mean ± standard deviation (SD) for continuous variables and number (%) for categorical variables. Data were compared with Mann Whitney U test or chi square test as appropriate and two tailed statistical tests were considered significant if p value was <0.05, two data sets were provided comparing demographic and secondary outcomes before and after matching.

Stata® was used for the propensity score matching (Stata Corp. 2017. Sta ta Statistical Software: Release 15. College Station, TX: Stata Corp LLC) and for the rest of the statistics SPSS® was used (IBM Corp. Released 2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp).

Results

During the year 2017 there were 2771 admissions to our ICU, 1740 patients were excluded permanently due to exclusion criteria that were not expected to revert, the remaining 1031 patients were screened daily for possible enrollment in the early mobilization program, 218 patients were enrolled in the program, whereas 813 patients were not and served as the control group (Figure 1).

Figure 1: Patients enrolment.

Demographic characteristics and secondary outcomes before and after matching are summarized in Table 2. Before matching, the control group had significantly fewer males (32.8%) as compared to the enrolled group (47.7%, p=0.0001), other demographics did not differ between groups. Similarly, there was a statistically significant difference between the average ICU LOS for the enrolled group of 15.8±8.2 and that of the control group of 26.6±12 (p=0.03), the enrolled group had an average day of weaning trials from mechanical ventilation of 4.1±2.6 which was also significantly lower than that of the control group of 8.2±18 (p=0.03), ICU mortality rate of enrolled patients of 13.3% was significantly lower compared to that of the control group of 21% (p=0.01).

Table 2: Demographics and secondary outcomes.

Propensity score matching resulted in 218 pairs with one enrolled and one un-enrolled patient in each pair, utilizing the variables with p value<0.2 in the logistic regression model, which was well fitted with a Hosmer Lemeshow p value of 0.07, those variables were: age, gender, APACHE 4 score, source of ICU admission and general diagnostic category (Table 3).

Table 3: Logistic regression model.

Propensity score matching reduced the mean bias in the treatment effect by 40% (Figure 2), as a result, any demographic imbalances between groups did not persist between matched groups, likewise, statistically significant difference in mortality did not persist (p=0.07, matched groups), whereas the significant difference in ICU LOS and weaning trials duration persisted (p values 0.04 and 0.03 respectively), we did not account in our data collection or analysis for adverse events (AEs ) taking place during ICU stay for any of the groups assuming that those rates would be similar since all patients were treated similarly.

Figure 2: Propensity score matching reduction of bias.

The primary outcome of effect of treatment was an estimated reduction of ICU LOS of 8.6 days (95% CI: 3.4-14.8, p=0.001).

Discussion

of our study provide statistical evidence that enrollment in the early mobilization program may significantly reduce LOS in ICU by 8.6 days, despite the fact that no causal relationship can be established based on statistics alone, these results clearly indicate that early mobilization of critically ill patients is associated with reduced LOS, similarly, as a secondary outcome in our study after matching average ICU LOS was significantly lower in the enrolled group (p=0.04) as well as the duration of weaning trials (p=0.03), ICU mortality, however, was not different between both matched groups (p=0.07) despite being numerically lower in the enrolled group. To our knowledge no previous study used propensity score matching to evaluate the impact of early mobilization on ICU LOS, however, this impact was studied by many researchers in different designs, several similar QI projects reported significant reduction of LOS for enrolled patients [7,14,15], whereas others failed to demonstrate that [16]. Likewise, reduction of median ICU LOS for the enrolled group was reported in a randomized controlled trial [1]. The aforementioned study in accordance with our study reported a significant difference in weaning days between both groups [1]. That study by Priyakshi et al. [1] as well as others report more ventilator free days for enrolled patients, which may correlate to fewer weaning trial days [17].

ICU mortality was not significantly different between matched groups, it is not a frequent measure in similar studies, however, a similar finding was reported by a controlled trial on stroke patients [18] and in a prospective trial the mortality rate of the enrolled patients was a very close to ours 18.8%, but there was no comparison group [19].

Regardless of the different measured outcomes of effectiveness, many published literature reviews and meta-analyses, although were not able to conclusively confirm benefit, yet they all concluded the feasibility, safety, potential beneficial effect and cost effectiveness of early mobilization of critically ill patients [2,9,19,20].

Conclusion

Early mobilization programs for critically ill patients may be associated with reduction of ICU LOS and weaning trial period.

Limitations

Although propensity score matching was designed to perform similar to randomized control trials (RCT) using retrospective data, yet it can’t have the same strength of an RCT as scientific evidence. The number of enrolled patients in our program (218) was lower than that reported in most studies on the topic, we did not account for AEs taking place during ICU stay which may prolong ICU LOS or affect outcomes, our report does not include any follow up measures after ICU discharge, such as hospital LOS, or qualitative measures of the patient’s lives after hospital discharge, such as return to work or level of assistance rating, finally, the variables used to generate propensity scores were few and general, when there are far more predictors of patient’s outcome in ICU.

Acknowledgment

The authors would like to thank all ICU staff at King Saud Medical City for their dedicated efforts to care for their patients and their cooperation conducting this study.

References

1. Priyakshi B, Narasimman S, Cherishma D, Shabari K (2012) Effect of graded early mobilization versus routine physiotherapy on the length of intensive care unit stay in mechanically ventilated patients: A randomized controlled study. Int J Health Allied Sci 1: 172-177.
2. Hunter A, Johnson L, Coustasse A (2014) Reduction of intensive care unit length of stay: The case of early mobilization. Health Care Manag 33: 128-135.
3. Gosselink R, Bott J, Johnson M (2008) Physiotherapy for adult patients with critical illness. Recommendation of European respiratory society and European society of intensive care medicine task force on physiotherapy for critically ill patients. Intensive Care Med 34: 1188-1199.
4. Brummel N, Jackson J, Girard T, Pandharipande PP, Schiro E, et al. (2012) A combined early cognitive and physical rehabilitation program for people who are critically ill: The activity and cognitive therapy in the intensive care unit (ACT-ICU) trial. Phys Ther 92: 1580-1592.
5. Lipshutz AK, Gropper MA (2013) Acquired neuromuscular weakness and early mobilization in the intensive care unit. Anesthesiology 118: 202-215.
6. Shorr AF (2002) An update on cost-effectiveness analysis in critical care. Curr Opin Crit Care 8: 337-343.
7. Needham DM, Wang W, Desai SV, Mendez-Tellez PA, Dennison CR, et al. (2007) Intensive care unit exposures for long-term outcomes research: Development and description of exposures for 150 patients with acute lung injury. J Crit Care 22: 275-284.
8. Needham DM, Wang W, Desai SV, Mendez-Tellez PA, Dennison CR, et al. (2007) Intensive care unit exposures for long-term outcomes research: Development and description of exposures for 150 patients with acute lung injury. J Crit Care 22: 275-284.
9. Bailey P, Thomsen GE, Spuhler VJ, Blair R, Jewkes J, et al. (2007) Early activity is feasible and safe in respiratory failure patients. Crit Care Med 35: 139-145.
10. Gensheng Z, Kai Z, Wei C, Yucai H, Zhongheng Z (2018) The effect of early mobilization for critical ill patients requiring mechanical ventilation: A systematic review and meta-analysis. J Emerg Crit Care Med 2: 9.
11. Carney N, Totten AM, O'Reilly C, Ullman JS, Hawryluk GW, et al. (2017) Guidelines for the management of severe traumatic brain injury, fourth edition. Neurosurgery 80: 6-15.
12. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, et al. (2017) Surviving sepsis campaign: International guidelines for management of sepsis and septic shock: 2016. Intensive Care Med 43: 304-377.
13. Brookhart MA, Wyss R, Layton JB, Stürmer T (2013) Propensity score methods for confounding control in non-experimental research. Circ Cardiovasc Qual Outcomes 6: 604-611.
14. Gi HK, Jung WC, Jong HL, Kyeong SO, Eun SJ, et al. (2015) Effect of vitamin E in non-alcoholic fatty liver disease with metabolic syndrome: A propensity score-matched cohort study. Clin Mol Hepatol 21: 379-386.
15. Engel HJ, Tatebe S, Alonzo PB, Mustille RL, Rivera MJ (2013) Physical therapist-established intensive care unit early mobilization program: Quality improvement project for critical care at the University of California San Francisco medical center. Phys Ther 93: 975-985.
16. McWilliams D, Weblin J, Atkins G, Bion J, Williams J, et al. (2015) Enhancing rehabilitation of mechanically ventilated patients in the intensive care unit: A quality improvement project. J Crit Care 30: 13-18.
17. Clark DE, Lowman JD, Griffin RL, Matthews HM, Reiff DA (2013) Effectiveness of an early mobilization protocol in a trauma and burns intensive care unit: A retrospective cohort study. Phys Ther 93: 186-196.
18. Brahmbhatt N, Murugan R, Milbrandt EB (2010) Early mobilization improves functional outcomes in critically ill patients. Crit Care 14: 321.
19. Tay-Teo K, Moodie M, Bernhardt J, Thrift AG, Collier J, et al. (2008) Economic evaluation alongside a phase II, multi-centre, randomised controlled trial of very early rehabilitation after stroke (AVERT). Cerebrovasc Dis 26: 475-481.
20. Hashem MD, Parker AM, Needham DM (2016) Early mobilization and rehabilitation of patients who are critically ill. Chest 150: 722-731.
21. Kalisch BJ, Lee S, Dabney BW (2014) Outcomes of inpatient mobilization: A literature review. J Clin Nurs 23: 1486-1501.