The Italian version of the Postural Assessment Scale for Stroke Patients (PASS): transcultural translation and validation

Introduction

Stroke is currently one of the most common causes of disability and dependence among the older adult population in developed countries (1-3). In the European Union, there are over 9.5 million stroke survivors and a 27% increase is expected in the next three decades, due to population aging and improved survival rates (4).

In addition to compromising the limb mobility in the affected side, a stroke causes an alteration of postural control and balance (5, 6). The ability to maintain balance in the sitting position, standing, and in postural variations is essential for the recovery of independence, and the close correlation between postural control of the trunk in the acute phase and future functional ability is recognized (6-10).

Being able to predict the degree of recovery at an early stage after the stroke onset allows the medical and rehabilitation team to optimize time, tools, and resources in planning goals and treatment (7, 8, 11). Hence the need to identify valid and adequate assessment tools. The Trunk Control Test (TCT) (12, 13) is probably the most used and feasible in the acute stage. Of the four scales for trunk control compared in 2019 by Fil Balkan et al (14), the TCT was found to be the most time-efficient and with a better predictive value, but showed a floor effect. In clinical practice a ceiling effect is also frequently observed within the first weeks (15). Other scales proposed in the literature, for example, the Fugl-Meyer (16, 17), require a long time and demanding training for their use (18). Others, such as the Trunk Impairment Scale (TIS) (5), evaluate trunk control only in a sitting position.

The Postural Assessment Scale for Stroke Patients (PASS), conceived by Benaim et al and published in 1999 (19), evaluates the ability to maintain stable postures and balance during positional changes. It can be applied to all patients with stroke, even those with minimal postural control, in the first 3 months. The validation studies confirmed the structural validity of the PASS, excellent inter- and intra-operator reliability, high internal consistency (19), and the absence of floor/ceiling (F/C) effects when applied to the target population in the first weeks post-stroke (19, 20). A ceiling effect has been found for patients with high functional ability (21, 22).

Recent studies showed that, compared to the Berg Balance Scale, the PASS is better able to detect balance improvements in patients with severe balance deficits (23), and that it is a valid instrument to assess balance at an early stage (20) but also in the subacute and chronic phase (24, 25). It is an excellent early predictor for autonomy in both basic activities of daily living (ADLs) and instrumental ADLs (IADLs) (7, 19, 26), consistent with the results detected at 3 months with the Functional Independence Measure (FIM) (19) or at 6 months with the Barthel Index (BI) and the Frenchay Activities Index (19, 27). The predictive power is greater than the Fugl-Meyer’s (7), even in foreseeing the patient’s walking ability after discharge (28). Because of its properties and short evaluation time, the PASS is used worldwide and validated versions have been produced in Portuguese (29), Swedish (30), Spanish (31), Norwegian (32), and Turkish (33).

Moreover, the PASS is among the 14 highly recommended outcome measures selected in 2013 by the American Physical Therapy Association for individuals with stroke (34).

It is therefore advisable to use this tool in the clinical context. Although the PASS is commonly used in Italy, there is no cross-culturally validated version in this language. This study aimed to provide an Italian version of the PASS, going through a cross-cultural validation process, assessing its validity and intra-/inter-rater reliability.

Methods

The study was conducted in two phases (Fig. 1): (1) translation and cross-cultural adaptation of the first version of the Italian PASS (PASS-IT), followed by a pilot study to resolve possible critical issues; (2) a psychometric evaluation, assessing the validity and reliability of the PASS-IT.

Results

Psychometric evaluation

Fifty-one patients had been initially recruited; two were excluded at a later time because they were also diagnosed with a brain tumor. The characteristics of the participants are presented in Table 1.

Further information about participants’ sensory disorders, unilateral spatial neglect (USN), upper/lower limb spasticity and function, as well as frequency distributions and percentage of the scores collected for each item are provided in the Supplementary Material (Table S1 and S2, respectively).

		Validity and intra-rater reliability study N = 49	Inter-rater reliability study N = 30
Age (in years)	Mean ± SD	71.96 ± 12.46	68.57 ± 13.02
	Median (range)	75 (41-89)	69 (41-89)
Gender	Women	21 (42.86%)	14 (46.67%)
	Men	28 (57.14%)	16 (53.33%)
Diagnosis	Ischemic	40 (81.63%)	24 (80%)
	Hemorrhagic	9 (18.37%)	6 (20%)
Brain injuries	Right	24 (48.98%)	14 (46.67%)
	Left	25 (51.02%)	16 (53.33%)
Days between date of ictus and date of testing	Mean ± SD	5 ± 2.68	5 ± 2.59
	Median (range)	4 (1-11)	5 (1-12)

SD = standard deviation.
N = 49 patients included in the validity and intra-rater reliability study;
N = 30 patients included in the inter-rater reliability study.

A principal component analysis was performed. The KMO test (0.880, p < 0.01) confirmed the appropriateness of the factor analysis and of the sample size. Bartlett’s test of sphericity was < 0.001. The scree plot (Figure S1 in Supplementary Material), the overall variance, and the pattern matrix showed a two-dimensional structure. The two components were studied with a Varimax rotation with Kaiser normalization (Table S3 in Supplementary Material). In factor analysis, items are allocated to factors according to the highest factor loadings, typically using a threshold of 0.3 or 0.4. Two unexpected groups of items were identified: items 1-4 plus 12 (activities performed in lying/sitting position, “non-weight-bearing”) and items 5-11 (activities standing, “weight-bearing”). High reliability was found for both subsections: “non-weight-bearing” (ICC = 0.865; 95% CI: 0.795-0.917) and “weight-bearing” (ICC = 0.949; 95% CI: 0.924-0.968). These two subsections, which were not further investigated in this study, do not coincide with those of the original PASS (“maintaining a posture” and “changing posture”).

Table 2 shows the scores for the four scales administered by Researcher 1 on the same test occasion: PASS-IT, TCT, BI, and FAC. Spearman’s rho (ρ) showed high concurrent validity between the PASS-IT and the TCT (ρ = 0.845, p < 0.001) and a strong correlation with the BI (ρ = 0.884, p < 0.001) and the FAC (ρ = 0.889, p < 0.001).

	Median (IQR)	Minimum	Maximum	Concurrent validity/correlation with the PASS-IT
PASS-IT	27 (18)	3	36	---
TCT	100 (38)	0	100	ρ = 0.845, p < 0.001
BI	45 (80)	0	100	ρ = 0.884, p < 0.001
FAC	1 (4)	0	5	ρ = 0.889, p < 0.001

BI = Barthel index; FAC = Functional Ambulation Categories; IQR = interquartile range; PASS-IT = Postural Assessment Scale for Stroke Patients, Italian; TCT = Trunk Control Test.

Table 3 shows the internal consistency results. For each item the median score with Interquartile Range (IQR) is shown, together with the item-to-total correlation, corrected inter-item correlation, and Cronbach’s α when the item is deleted. The item-to-total correlation shows a value of 0.390 for the first item, while the others range from 0.663 to 0.939. The corrected inter-item correlation shows a value of 0.363 for the first item, with the others ranging from 0.617 to 0.921. Cronbach’s α coefficient, regardless of which item is deleted, is always > 0.90 (range 0.929-0.947).

	Intra-rater
	Median (IQR)	Item-to-total correlation	Corrected inter-item correlation	Cronbach’s α when item is deleted
1. Supine to affected side lateral	3 (0)	0.390	0.363	0.947
2. Supine to the nonaffected side lateral	3 (0)	0.663	0.628	0.941
3. Supine to sitting up on the edge of the table	3 (1)	0.800	0.771	0.937
4. Sitting without support	3 (0)	0.734	0.686	0.938
5. Sitting to standing up	3 (3)	0.939	0.921	0.929
6. Standing with support	3 (3)	0.922	0.897	0.930
7. Standing without support	3 (3)	0.933	0.911	0.930
8. Standing on nonparetic leg	0 (1)	0.703	0.643	0.940
9. Standing on paretic leg	0 (1)	0.680	0.617	0.940
10. Standing, picking up a pencil from the floor	2 (3)	0.882	0.843	0.933
11. Standing up to sitting down	3 (3)	0.929	0.907	0.930
12. Sitting on the edge of the table to supine	3 (1)	0.804	0.779	0.938

IQR = interquartile range.

In the intra-rater reliability study, the mean total score for the PASS-IT is 24.15 ± 10.14 for the first assessment and 24.29 ± 10.16 for the second assessment made by Researcher 1. The mean interval between assessments was 9h43’ (SD = 6h54’; range 3h30’ to 21h30’). There is high reliability between total scores (ICC = 0.942; 95% CI: 0.914-0.963; p < 0.001) and for each item between first and second assessment, with ρ ranging from 0.817 to 0.991.

In the inter-rater reliability study, the mean total score for the PASS-IT is 26.00 ± 9.60 for Researcher 1 and 26.03 ± 9.60 for Researcher 2. The assessments are highly consistent for total scores (ICC = 0.940; 95% CI: 0.903-0.968; p < 0.001) and for single items, with ρ ranging from 0.988 to 1.000.

The SEM is 1.72 points for the intra-rater, and 1.63 points for the inter-rater reliability. The MDC_95% is 4.76 based on intra-rater reliability data.

In the PASS-IT, five patients (10.20%) reached the maximum score (mean 24.22/36; range 3 to 36). Nine patients (18.37%) scored between 34 and 36/36. No one scored 0. In the TCT, 29 patients (59.18%) received the highest score (mean 77.63/100; range 0 to 100). The zero score was given to two patients (4.08%).

With the new sequence of items, the time of administration decreased (pilot study, mean 12′14″, range 5 to 20 minutes; psychometric study, mean: 7′55 ″, range 4′35 ″ to 15′25 ″).

Discussion

The PASS-IT, with concise operating instructions, showed a high intra- and inter-rater reliability, reflecting a uniform application and interpretation of the scale. The involvement of a large group of physiotherapists, together with the rigorous method followed for the cross-cultural translation, can be considered a strength of this study.

The modified sequence of the items, chosen also for the Swedish (Swe-PASS) (30) and Norwegian versions (Swe-PASS-NV) (32), allowed a quicker and smoother administration, avoiding repeated unnecessary changes of posture. The clear distinction between the original two sections of the scale (“Maintaining a posture” and “Changing posture”) is not featured, but the total score of the test does not change.

Although the PASS requires no training for use, the pilot study highlighted the need for operating instructions, consisting of concise and pragmatic indications, shown before each item in the evaluation form, aiming to increase the uniformity both in the administration (setting, instruction given to the patient) and in the interpretation of the results. For example, for items 4, 7, 8, 9 (sitting without support, standing without support, standing on the nonparetic/paretic leg) it is essential to use a stopwatch, as planned by the authors of the PASS (20). Underestimating its use in the pilot study led to imprecise scores.

In items 1-3, 5, 10-12, different scores are expected based on the amount of help received (“much help,” “little help,” “without help”) (19). There are no tools to quantify numerically the help provided, which is subjectively affected by age, technique, experience, build, and physical training of the operator. The Swe-PASS replaced “much help” with “support from two persons” and “little help” with “support from one person” (30). After extensively discussing this aspect, the original definition was preferred for the PASS-IT.

The original PASS describes a setting with a Bobath-type plane 50 cm in height, with the person’s feet resting on the floor. For short people, it is difficult to touch the ground from this height, except by moving dangerously forward on the edge of the couch. For taller people, excessive bending of the lower limbs leads to unfavorable leverage. A sitting position with hips and knees bent at 90° is therefore advisable, as recommended by the new instructions for items 4 and 5. When the patient rests on an anti-decubitus mattress, it is essential that the mattress be maximally inflated, to allow stability. Items in standing were tested wearing shoes, for greater safety and hygiene, even if the original PASS doesn’t specify whether the patient should be wearing shoes or be barefoot.

While other cross-cultural validation studies evaluated the patients at a chronic stage, or through a video-recorded performance (31, 32), our study was carried out in an acute setting, when a wide variety of factors (functional improvement, caution, fatigue, fear, disorientation) can produce sudden changes. For this reason, like in the Swedish validation study (30), we chose a short interval (< 24 hours) between the intra-rater observations, despite this representing a weakness of the study because of recall bias. A longer interval could result in higher grades at the second assessment, due to the functional recovery of the patient, or because posture changes had been practiced with the physiotherapist or the ward staff.

The value of verbal indications for carrying out the activity was also questioned. To standardize the application of the scale, it is important that no verbal indications be provided that are useful for its performance. Should they become necessary, the verbal indications would be considered as “little help,” like in the TCT (12).

The operating instructions emphasize that the PASS, for each of the 12 items, evaluates the ability to perform the activity, without considering its quality. They also draw attention to the fact that only one attempt is allowed for each item.

The PASS does not evaluate the ability to perform activities while seated. When this aspect is essential, it is advisable that more specific tools be used, such as the TIS (5), of which the validated translation in Italian is available (42), or the Function in Sitting Test (43).

No patient scored zero in item 1 (Supine to affected side lateral), and only one patient scored zero in item 2 (Supine to nonaffected side lateral). This can be partially explained by our exclusion criteria. Breistein et al (32) discussed the possibility that score 0 might be redundant for this item, since even patients with little to no functional independence can be turned on one side with help from one or two persons. We agreed that a zero be scored when medical conditions contraindicate rolling the patient to one side, but also when the patient shows no participation or involvement in the action.

In item 7, the highest score is given if the person “can stand without support for more than one minute and at the same time perform arm movement above the shoulder level.” It is important to select specific movements, with a predefined number of repetitions, to obtain comparable results. Persson et al (30) indicate the act of moving “hand/s from the forehead to the neck (like pulling your fingers through your hair)” and bringing the arm back, relaxed along the trunk, without however specifying the number of repetitions. We recommend at least five repetitions of this movement.

As shown in Table S2, at first assessment 59.2% of patients were not able to stand on the non-paretic leg (item 8), 63.3% were not able to stand on the paretic leg (item 9), and 46.9% were not able to pick up a pencil from the floor (item 10). These findings are in line with those of the Swe-PASS (30) and with the results described by Benaim et al (19) for the group of patients evaluated on day 30 after a stroke. In items 8 and 9 it is not clear whether the foot which is not bearing the weight must be lifted off the ground, or whether bending the hip and/or the knee is required. During the pilot study, these items had the worst inter-rater agreement. It was therefore specified that the foot must be lifted off the ground, with the opposite one entirely bearing the weight.

We consider the two-dimensional structure identified by the principal component analysis an unexpected and interesting finding. There is a clear division between items “non-weight-bearing” (1-4 plus 12), where the patients show much better performance, and the “weight-bearing” section (items 5-11), with a floor effect for some tasks. Nothing changes in the way the scale is administered and scored, but we believe that a “non-weight-bearing” and “weight-bearing” subdivision can be relevant to better comprehend the patient’s improvement and need for rehabilitation.

The present study confirmed a good correlation with the TCT (ρ = 0.845, p < 0.001) and the FAC (ρ = 0.889, p < 0.001), which had not been tested before. Our findings of a good correlation with the BI (ρ = 0.884, p < 0.001) are consistent with the values reported by Mao et al (20) and Chien et al (40).

The PASS-IT showed very high reliability, in line with Benaim et al (19) and other transcultural validations (30-33).

Considering that our patients were in the acute or early subacute phase after a stroke, and nine of them (18.37%) scored between 34 and 36/36, we agree with Chinsongkram et al (21) and with Wang et al (22) that a ceiling effect is plausible already in the first weeks for patients with high functional ability. The TCT showed a significant ceiling effect, with 29 patients (59.18%) earning the highest score. Considering this, the SEM (PASS = 1.72, normalized 4.77; TCT = 11.30), and the MDC (PASS = 4.76, normalized 13.22; TCT = 31.34), in our clinical practice the PASS-IT is more advisable, being more responsive to slight changes (e.g., going from “with much help” to “with little help”). Moreover, it evaluates the patients also in the standing position and single-leg stance.

The original PASS does not report an MDC value. Hsueh et al (44) indicate an MDC of 1.8 ± 1.7 for acute stroke, but their assessment was conducted at 14 and 30 days from onset. Breistein et al (32) calculated smallest detectable difference (SDD) = 1.9 points (intra-rater) and 2.7 points (inter-rater), but attested that “the measurement error may be considered to be artificially low” due to the use of video recorded assessment, without a real change in the participants’ performance.

Since different statistical methods were employed, a direct comparison of the results is often impossible.

Conclusion

The PASS-IT is a valid and reliable tool, suitable for clinical and physiotherapy practice in the acute and subacute stage. The final version of the cross-cultural translation, which includes short operating instructions and a different sequence for the 12 items, overcame critical issues encountered during the pilot study. It serves the purpose of promoting a uniform application and interpretation of the scale among Italian health professionals and researchers. Further study is needed to investigate the potential information provided by considering the scale in its two dimensions, “weight-bearing” and “non-weight-bearing.”

Acknowledgments

The authors wish to thank the patients who participated in this study and the physiotherapists of the Unit of Physical Medicine and Rehabilitation of the IRCCS Azienda Ospedaliero-Universitaria di Bologna for their help in thoroughly discussing every critical aspect, and for being such a committed and passionate work team.

A sincere appreciation goes to the reviewers who, with their competence and attention, provided valuable guidance.

Disclosures

Conflict of interest: The authors declare that there is no conflict of interest.

Financial support: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Ethics approval: The study was approved by the Ethical Committee of the IRCCS University Hospital of Bologna (Reference N. 216/2017/O/Sper).

Authors contribution: EL: Conceptualization, Data Curation, Formal Analysis, Investigation, Methodology, Project Administration, Resources, Supervision, Validation, Visualization, Writing – Original Draft, Writing – Review & Editing; NG: Conceptualization, Data Curation, Formal Analysis, Investigation, Methodology, Project Administration, Software, Supervision, Validation, Visualization, Writing – Original Draft, Writing – Review & Editing; AB: Conceptualization, Data Curation, Investigation, Project Administration, Supervision, Validation, Writing – Review & Editing; AB: Conceptualization, Investigation, Project Administration, Resources, Supervision, Validation; LG: Conceptualization, Data Curation, Formal Analysis, Methodology, Resources, Software, Supervision, Validation, Visualization, Writing – Original Draft, Writing – Review & Editing; MR: Conceptualization, Data Curation, Formal Analysis, Resources, Software, Validation, Visualization, Writing – Original Draft, Writing – Review & Editing; DG: Conceptualization, Data Curation, Formal Analysis, Methodology, Resources, Software, Supervision, Validation, Writing – Review & Editing.

Data Availability Statement: The data presented in this study and the printable Italian Version of the PASS are available on request from the corresponding author.