Evaluating physical functioning in critical care: considerations for clinical practice and research

The World Health Organization (WHO) ICF framework defines functioning as an umbrella term for the interaction between three distinct constructs: body function and structure (physiological and anatomical structure of the body systems), activities (execution of a specific task within a standardized environment), and participation (involvement in everyday life situations) [15]. The ICF framework explicitly recognizes that functioning is affected by the interplay between an individual’s health condition and contextual factors, which may include personal (e.g., education) and environmental/social (e.g., home set-up, family support) factors [15].

Using this framework, physical functioning can be evaluated across the three ICF constructs. First, functioning can be evaluated in terms of physiological impairment at the level of individual organs or body systems (i.e., the “body function” level of assessment of the ICF) [15], with a specific focus on the neurological, cardiac, respiratory, and musculoskeletal systems. Second, functioning can be evaluated in terms of performance-based measurement focused on limitations in specific activities, such as sitting, standing, or walking [15]. Third, evaluation can include assessment of participation restrictions, such as the ability to perform activities of daily living (ADLs). These perspectives evaluate distinct aspects of physical functioning and thus structure and function impairment (e.g., muscle weakness) does not necessarily strongly correlate with activity limitations (e.g., 6-min walk test) and participation restrictions (e.g., ADLs) [13, 16, 17].

While post-ICU impairments in physical functioning are common, our understanding of the specific subgroups of patients at highest risk for such impairments, and with the greatest potential benefit from rehabilitation interventions, is evolving. Measuring physical functioning early and longitudinally in the ICU is important to identify patients at risk of poor physical outcomes, monitor intervention efficacy, and inform recovery trajectories [12, 18, 19].

Pre-ICU factors, such as age, comorbidities, and pre-ICU trajectories for muscle mass and physical functioning, impact on the physical functioning of patients in the ICU (Fig. 1). In addition, there are many factors related to critical illness and the ICU environment that can impact on impairment in physiological body systems that are critical to the physical functioning of patients in the ICU (Fig. 1).

The evaluation of physical functioning is complex and is influenced by multiple interacting factors, including strength, range of motion, proprioception, balance, cognition, and psychological issues (e.g., motivation) [20]. There are also unique patient and environmental factors (e.g., sedation, severity of illness, medical devices) specific to the ICU. Determining the specific purpose for assessing physical functioning is important when selecting an appropriate instrument. For example, if the purpose is to evaluate intervention efficacy, users should consider the specific effect of the intervention and match it with an instrument that evaluates that effect. Table 3 highlights that there are important differences when each physical function instrument is mapped to the relevant subdomains of the ICF framework. For example, the Chelsea Critical Care Physical Assessment Tool assesses both respiratory and mobility ICF subdomains; the ICU Mobility Scale only evaluates mobility subdomains; and the Physical Functional in ICU Test-scored is a composite measure of mobility, strength, and endurance. Hence, if the primary aim of an intervention is to improve patient mobility via increased muscle strength, it may be most appropriate to use a composite instrument which evaluates mobility and strength (e.g., Physical Functional in ICU Test-scored or Chelsea Critical Care Physical Assessment Tool) or separate instruments individually focused on strength and mobility (e.g., dynamometry, plus ICU Mobility Scale or Functional Status Score for the ICU). Whilst domains such as climbing and jumping, which are evaluated within the Acute Care Index of Function and the Critical Care Functional Rehabilitation Outcome Measure, are less relevant during an ICU admission, they are relevant later in the recovery process. Assessment of ‘climbing’ or stair walking ability is often a critical consideration in evaluating a patient’s safety for discharge to home. Currently there is not a single measure available that can be utilized across the entire recovery trajectory. Therefore, consideration of the elements evaluated under the subdomains of the ICF framework is important when selecting the relevant instrument based on the assessment purpose.

Relevant measurement properties to consider when selecting an instrument include the ability to measure what is intended (validity). This includes subjective interpretation (face validity), whether the instrument’s content adequately reflects the parameter of interest (content validity), comparison with other tools measuring a similar construct (construct validity), and prediction of future outcomes (predictive validity) [21, 22]. In addition, the ability to obtain accurate results within or between assessors (intra- and inter-rater reliability, respectively), or when measures are repeated longitudinally (test-retest reliability) is important. Instruments should detect change over time (responsiveness) and have a limited floor (proportion of patients scoring the lowest score possible) and ceiling (proportion of patients scoring the highest score possible) effect across the expected evaluation time points [22].

Notably, caution is needed if trying to extrapolate instruments developed for one setting or patient population to the ICU setting. This issue is particularly important given many unique issues within the ICU, including sedation, delirium, fatigability, and weakness, that can affect patient performance. For example, Acute Care Index of Function, De Morton Mobility Index, and Short Physical Performance Battery were initially developed for use in non-ICU patient populations (e.g., geriatrics and neurology), [23‐25] with relatively little evaluation, at present, within the ICU setting.

Table 1 and Additional file 1 (Table S1) synthesize data on measurement properties for each instrument. An adequate level of inter-rater reliability has been demonstrated for all measurement instruments, except the Short Physical Performance Battery, although this instrument has established reliability in geriatrics. All instruments have evidence of construct validity compared to other concurrent measurements of physical function and/or strength, except the Critical Care Functional Rehabilitation Outcome Measure. This instrument only has published data on face/content validity. Seven measurement instruments have evidence of predictive validity. The most commonly evaluated predictive outcome was discharge to home, which was evaluated in six instruments. The Short Physical Performance Battery was the only instrument not predictive of discharge to home, albeit the study may have been under-powered for this assessment [26]. Accurately predicting individuals unable to be discharged to home is important to help optimize the consistency, appropriateness, and timeliness of discharge planning recommendations and rehabilitation referral. There has been limited evaluation of the predictive validity beyond hospital discharge with only three instruments (ICU Mobility scale, Surgical Optimal Mobility scale, and Physical Functional in ICU Test-scored) examining post-hospital mortality with conflicting findings (Table 1; Additional file 1: Table S1).

The ability to detect a clinically meaningful change over time (responsiveness) was examined in five instruments (Table 1; Additional file 1: Table S1): the Chelsea Critical Care Physical Assessment Tool; the Functional Status Score for the ICU; the ICU Mobility Scale; the Physical Functional in ICU Test-scored; and the Short Physical Performance Battery. All demonstrated significant change over time within the ICU, and moderate to large effect sizes (an indicator of moderate to good responsiveness) were observed for the Functional Status Score for the ICU and Physical Functional in ICU Test-scored.

The presence of floor and ceiling effects are important considerations in assessing the recovery trajectories of patients and the intervention efficacy [22]. High floor or ceiling effects indicate that the instrument is too challenging or too easy, respectively, limiting its ability to detect a change in the physical functioning of patients. The majority of instruments have low floor and ceiling effects during an ICU stay (Table 1; Additional file 1: Table S1). However, the Short Physical Performance Battery demonstrated large floor effects which limits its potential utility in the ICU (Table 1; Additional file 1: Table S1).

Based on published measurement properties alone, the most robust ICU instruments are: Physical Functional in ICU Test-scored; Chelsea Critical Care Physical Assessment Tool; Functional Status Score for the ICU; and ICU Mobility Scale (Table 1). Ongoing research is needed to further understand the measurement properties of existing instruments to ensure appropriateness and usability within the ICU setting.

All instruments outlined herein (Tables 1 and 2) are dependent on patient effort. Consequently, assessing the feasibility of each instrument’s use within the ICU is critical. Feasibility should consider the requirements of the instrument, including issues related to a patient’s alertness, ability to follow instructions, motivation, weakness, and fatigability. A standardized method for determining patient mental capacity (including validated and reliable determinations of pain, sedation, and delirium status) is important to enable comparison of results across patients [27] (Fig. 3). The Perme ICU Mobility Score is unique as it includes evaluation of potential barriers to mobility that may affect patient performance (e.g., medical devices, pain, and respiratory support). Impairments in the balance of patients may also affect performance, with the De Morton Mobility Index and Short Physical Performance Battery including balance evaluation.

The levels of expertise, training, and time required, as well as any specialized equipment, are important in assessing clinical utility. All instruments require minimal additional equipment, apart from the Chelsea Critical Care Physical Assessment Tool which requires a handgrip dynamometer, and the Acute Care Index of Function which requires a set of five steps (Table 2). Dedicated ICU training packages are available for three instruments: Chelsea Critical Care Physical Assessment Tool; Functional Status Score for the ICU; and Physical Function in ICU test-scored (Table 2). The fastest tests are the simple one-item mobility scales that indicate the patients highest level, while other more comprehensive instruments require more time to assess multiple specific activities and/or levels of assistance required (Table 2).

We propose a staged approach for assessing physical functioning in the ICU (Fig. 3). In terms of pre-ICU status, we recommend obtaining physical functioning data as part of the patient history to inform appropriate patient goals for recovery and rehabilitation [2]. The ability to obtain a validated baseline measure of physical functioning (or pre-ICU health status) is challenging due to the severity of illness, sedation, and reduced ability of patients to engage in volitional assessments. The Clinical Frailty Scale can be used to obtain a baseline assessment of frailty. Patients who are frail prior to ICU admission have worse mortality and morbidity, and require institutionalization at discharge; thus, frailty may be a useful prognostic tool [28]. Similar to the process used with the Functional Independence Measure instrument, commonly used throughout the inpatient rehabilitation setting, it is possible to conduct a baseline assessment for the Functional Status Score for the ICU measure via proxy assessment, as performed in prior research [29]; however, this baseline version of the Functional Status Score for the ICU has not been specifically validated. The ability to measure pre-ICU physical functioning is an area for future research.

Screening for mental capacity should commence from ICU admission and include assessments of pain, sedation, and delirium status [27]. We also recommend regular screening for muscle weakness using the Medical Research Council sum-score. It is likely less important to evaluate physical functioning in ICU patients who lack muscle weakness; however, strength should not be a sole guide for determining the need for physical functioning assessment because strength and function are only weakly correlated in ICU survivors [16]. At present, there is a lack of robust, validated predictive models for physical functioning impairments within the ICU. There is a predictive model for physical functioning after hospital discharge, but not whilst in hospital [12]. Therefore, identification of patients who need evaluation of physical functioning in the ICU is largely reliant on clinical judgment regarding many potential risk factors (Fig. 1). Once the patient can follow commands, we recommend, at a minimum, one of the four recommended physical functioning tools: Physical Functional in ICU Test-scored; Chelsea Critical Care Physical Assessment Tool; Functional Status Score for the ICU; and ICU Mobility scale. Summary information about these instruments (including how to access and use them) is available through a free website: www.​improveLTO.​com.

When selecting specific instrument(s) for a particular ICU setting, the following are important considerations: available clinician resources and expertise; and rationale for assessment (e.g., simple versus comprehensive evaluation). In settings with limited access to rehabilitation clinicians, a simple one-item scale (e.g., ICU mobility scale) can be used, which can be feasibly completed by the bedside ICU nurse. For patients with identified mobility restrictions, consultation with physiotherapists and occupational therapists may be warranted, with more comprehensive instruments used as part of their routine clinical evaluation (Fig. 3).