Fear of falling in acute stroke: The Fall Study of Gothenburg (FallsGOT)

ABSTRACT

Objective Little is known about which factors are associated with a patient’s fear of falling (FoF) after acute stroke. The aim of this study was to investigate baseline variables and their association with FoF during rehabilitation in acute stroke.

Patients and methods The study population consisted of the 462 patients with acute stroke who were admitted to a stroke unit, included in the observational study “The Fall Study in Gothenburg (FallsGOT)” and were able to answer a single question: “Are you afraid of falling?” (Yes/No). To analyze any association between FoF and clinical variables, univariable and multivariable stepwise multiple logistic regression analyses were performed.

Results In the stepwise multivariable regression analysis, only female sex (OR = 2.25 [95% confidence interval (CI) 1.46–3.46, p = 0.0002]), the use of a walking aid (OR 3.40, [95% CI 2.12–5.43, p < 0.0001]), and postural control as assessed with the SwePASS total score were statistically significant associated with FoF. Among patients with a SwePASS score of 24 or less, the OR was 9.41 [95% CI 5.13–17.25, p < 0.0001] for FoF compared to patients with a SwePASS score of 31 or above; among the patients with a SwePASS score of 25–30, the OR was 2.29 [95% CI = 1.36–3.83, p = 0.0017].

Conclusions Our findings provide valuable insight for those involved in stroke rehabilitation during the acute phase after stroke. FoF is associated with poor postural control, female sex and the use of a walking aid.

KEYWORDS:

• Acute stroke
• fear of falling
• postural balance
• rehabilitation

Introduction

Stroke may cause physical impairments such as weakness, paralysis, sensory disturbances, and impaired postural control¹; it can also cause mental fatigue, depression and impaired cognitive function.² Both physical and mental impairments can contribute to a fall,³ a common complication after a stroke. Among people who have survived a stroke, 22–48% have experienced at least one fall during a hospital stay^3–5 or a later stage of recovery.^6,7 A fall, according to the World Health Organization, is defined as “an event which results in a person coming to rest inadvertently on the ground or floor or other lower level, with or without injury”.⁸ Compared with age-matched controls, people who have experienced a stroke are significantly more likely to report fear of falling (FoF).⁶ FoF can be defined as low confidence in one’s ability to carry out everyday activities while maintaining postural control without falling.⁹ There is a reported prevalence of 32–83% for FoF between the first six months and just over four years after stroke onset.^6,10,11 The prevalence of FoF is considerably higher among women (62–78%) than among men (18–38%).^12–14

FoF can lead to uncertainty, inactivity, reduced physical function, affected social relationships and impaired quality of life.^15–17 Furthermore, studies have revealed an association between FoF and earlier falls^18–20 and between FoF and activity after stroke, such as balance and functional mobility.²⁰ In addition, FoF has been linked to anxiety and depression^11,21,22 and has been shown to affect the patient’s level of activity and participation.²³

To optimize the rehabilitation from start, it seems important to early identify those at risk of developing FoF. However, previous studies have primarily focused on patients at a later stage of stroke recovery,^{6,11,14,19,20,23–26} while studies on FoF during the first days in the hospital after stroke are limited.^27–29 One small study that included 28 patients with a first-ever stroke and an average age of 59 years showed that more than 50% experienced FoF before hospital discharge.²⁷ Another study analyzed a possible correlation between FoF, measured using the Falls Efficacy Scale, Swedish version, and the patient’s actual risk of falling in the acute phase after stroke, assessed using the Berg Balance Scale (BBS) and the Timed Up & Go test.²⁸ However, no clear correlation was found.²⁸ Recently, The Fall Study in Gothenburg (FallsGOT),²⁹ found that 51% of participants reported FoF within four days after admission to the stroke unit. In addition, as previously presented in the FallsGOT, 74% of the patients who fell during hospitalization reported FoF compared with 48% of patients who did not fall.²⁹ This difference in the reported FoF among patients who did and did not fall during hospital stay was statistically significant in a univariable analysis, but not in a multivariable analysis, when other factors were taken into account.²⁹ The aim of this study was to investigate baseline variables and their association with FoF during rehabilitation in acute stroke.

Patients and methods

The study is part of the prospective and observational FallsGOT²⁹ which primary aimed to investigate factors associated with falling after stroke, while the present study focuses on FoF. The study was approved by the Regional Ethics Committee in Gothenburg 25 February 2014 (ref 004-14) and has previously been described in detail elsewhere.²⁹ The reporting of this study conforms with the Strengthening of Reporting of Observational Studies in Epidemiology (STROBE) Guidelines.³⁰ Study enrollment was consecutively conducted between 1 October 2014 and 30 June 2016. The inclusion criteria were patients aged 18 years or older with a diagnosis of a first-ever or recurrent clinical stroke who were admitted to the stroke unit of the Sahlgrenska University Hospital (SU)/Östra, Gothenburg, Sweden and were able to answer a question regarding FoF. Patients receiving palliative care were not considered for inclusion. Each patient gave written informed consent or, if he or she was unable to read or understand the patient information, informed consent was given by the next of kin.

The dependent variable was FoF. FoF was defined as a positive answer to the question “Are you afraid of falling?” (Yes/No) asked by the physiotherapist. FoF was assessed in patients with stroke within four days after admission to the stroke unit; the admission day was counted as day zero. The current study is referring to the timepoint “acute stroke” since the first week post stroke has been defined as the acute phase after stroke³¹ and since it involves the time when FoF was assessed. The physiotherapist also assessed the patient’s postural control using the Swedish modified version of the Postural Assessment Scale for Stroke (PASS), SwePASS.³² The SwePASS consists of 12 items for which the physiotherapist assesses the patient’s postural control on an ordinal scale from 0 to 3. The total score ranges from 0 to 36 points, with higher scores indicating better postural control.^7,32,33 Based on previous evidence associated with the risk of falling, the SwePASS was divided into three categories (poor postural control: SwePASS-score ≤24, moderate postural control: SwePASS-score 25–30, and good postural control: SwePASS-score ≥ 31).²⁹ The SwePASS was used since it has shown adequate psychometric properties and ability to predict the risk of falling the first year after stroke⁷ and required only 8 min to be performed.³²

The physiotherapist also asked the patient about his/her previous level of physical activity based on the Saltin-Grimby Physical Activity Level Scale (SGPALS).³⁴ The SGPALS is a four-level ordinal scale in which patients rate their physical activity level from 1 to 4, where a higher score represents a higher level of physical activity.^35,36 Moreover, during the first four days of hospital stay, in median one day after admission, the physiotherapist and the occupational therapist respectively registered the patients’ possible use of a walking aid and/or a wheelchair.

Stroke severity was assessed using the Swedish version of the National Institutes of Health Stroke Scale (NIHSS)³⁷ as soon as possible after admission to the stroke unit. The NIHSS total score ranges from 0 to 42. Higher scores represent greater stroke severity: <5 points (mild stroke), 5–14 points (mild to moderately severe stroke), 15–24 points (severe stroke) and >25 points (very severe stroke).³⁷ Data concerning sex, age, stroke severity, length of hospital stay, stroke subtype, stroke localization, and the presence of risk factors and co-morbidity were collected from the patients’ medical record.

Statistical analysis

For descriptive purposes, numbers and percentages/median and interquartile range (IQR) are given for categorical variables, and mean, standard deviation, median, range, and IQR are provided where applicable for continuous variables. The univariable association between selected variables and FoF was analyzed using logistic regression. The significant variables were then included in a stepwise logistic regression from which a significant multivariable model was selected. No imputation of missing data was performed, as only 1/462 patients was excluded in the multivariable model due to missing data. Odds ratios (OR) with 95% confidence intervals (CIs) and p-values are presented. The goodness of fit for the multivariable logistic model was tested using the Hosmer and Lemeshow test and was satisfied (p = 0.22). All tests were two-tailed and conducted at 0.05 significance level. All analyses were performed using SAS software version 9.4 (Cary, NC, USA). Sample size was calculated in order to analyze risk factors for falling in the original study.²⁹

Results

Of the 504 patients included in “The Fall Study of Gothenburg”,²⁹ 42 were excluded due to their inability to answer the question: “Are you afraid of falling?” Thus, the study population in the present study consists of 462 patients. Of these, 338 (73.2%) were diagnosed with a first-ever clinical stroke, while 124 patients had a recurrent stroke. The 42 patients who were excluded from the study due to their inability to answer the question about FoF had more severe stroke according to the NIHSS: median NIHSS score of 7 (IQR 2–14, mean 8.7 ± 8.4) compared with 1 (IQR: 0–3, mean 2.7 ± 3.7) for the general study population. Assessment of FoF was performed a mean of 1.8 days and a median of 1 day after admission.

Table 1 presents baseline demographics and clinical data, while table 2 shows the prevalence of risk factors for stroke and co-morbidities. Table 3 demonstrates the odds ratios (OR) with 95% confidence intervals (95% CI) for the baseline variables as predictors of FoF based on the univariable and the multivariable logistic regression analyses. Older age, female sex, long hospital stay, poor postural control, higher NIHSS-score, the use of a walking aid, atrial fibrillation and the level of physical activity (SGPALS) were all statistically significant associated with FoF in the univariable analysis. However, in the multivariable regression analysis, only female sex (OR = 2.25 [95% confidence interval (CI) 1.46–3.46, p = 0.0002]), the use of a walking aid (OR 3.40, [95% CI 2.12–5.43, p < 0.0001]) and postural control, as assessed using the SwePASS total score, were statistically significant associated with FoF. Among the patients with a SwePASS total score of 24 or less, the OR was 9.41 [95% CI 5.13–17.25, p < 0.0001], and among the patients with a SwePASS score of 25–30, the OR was 2.29 [95% CI = 1.36–3.83, p = 0.0017] for FoF compared with patients with a SwePASS total score of 31 or above. The area under the receiver operating characteristic (ROC) curve, with 95% CI for the multivariable model, was 0.80 (0.76–0.84). In a logistic regression, the interaction between sex, FoF and falling was non-significant (p = 0.58; not shown in table).

Table 1. Baseline demographics and clinical data for the participants.

Characteristic	Total	FoF	No FoF
	n = 462	n = 237 (51.3%)	n = 225 (48.7%)
Age
Mean (SD)	74.8 (12)	76.9 (11.6)	72.6 (12.1)
Sex
Female, n (%)	226 (48.9)	141 (59.5)	85 (37.8)
Male, n (%)	236 (51.1)	96 (40.5)	140 (62.2)
NIHSS (n = 430)
Mean (SD)	2.7 (3.7)	3.3 (4.3)	2.1 (2.8)
Length of hospital stay (days)
Mean (SD)	12.2 (11.5)	15 (13.1)	9.2 (8.5)
Type of stroke, n (%)
Ischemic	431 (93.3)	224 (94.5)	207 (92)
Hemorrhagic	29 (6.3)	13 (5.5)	16 (7.1)
Subarachnoid	2 (0.4)	0 (0)	2 (0.9)
Stroke location (n = 449)
Right, n (%)	254 (55)	141 (59.5)	113 (50.2)
Left	186 (40.3)	86 (36.3)	100 (44.4)
Bilateral	9 (1.9)	4 (1.7)	5 (2.3)
Unknown	13 (2.8)	6 (2.5)	7 (3.1)
SwePASS total score (n = 461)
Mean (SD)	26.9 (8.4)	23.5 (8.8)	30.5 (6.2)
Median (IQR)	30 (24; 32)	26.5 (19; 30)	32 (30; 35)
Using a walking aid, n (%)	195 (42.2)	141 (59.5)	54 (24)

Notes:

FoF: fear of falling; SD: standard deviation; IQR: interquartile range; NIHSS: National Institutes of Health Stroke Scale, within 24 h of admission; SwePASS: the Swedish modified version of the Postural Assessment Scale for Stroke Patients, assessed within 4 days of admission.

Table 2. The prevalence of risk factors for stroke and co-morbidities among all patients and patients with and without fear of falling at baseline.

Characteristic	Total	FoF	No FoF
n (%)	462	237 (51.3)	225 (48.7)
Hypertension	336 (72.7)	177 (74.7)	159 (70.7)
Atrial fibrillation	123 (26.6)	73 (30.8)	50 (22.2)
Heart failure	49 (10.6)	27 (11.4)	22 (9.8)
Ischemic heart disease	98 (21.2)	53 (22.4)	45 (20)
Diabetes mellitus	97 (21)	51 (21.5)	46 (20.4)
Hyperlipidaemia	348 (75.3)	170 (71.7)	178 (79.1)
Smoking (n = 428)
Previous smoker	103 (24.1)	49 (22.3)	54 (26)
Current smoker	93 (21.7)	43 (19.5)	50 (24)
SGPALS (n = 459)
1 Physically inactive	253 (55.1)	161 (67.9)	92 (41.4)
2 Some light PA	186 (40.5)	69 (29.1)	117 (52.7)
3 Regular PA/training	20 (4.4)	7 (3)	13 (5.9)

Notes:

For categorical variables, n (%) is presented. FoF: fear of falling; SGPALS: Saltin-Grimby Physical Activity Level Scale; PA: physical activity.

Table 3. Univariable and multivariable logistic regression showing odds ratios for baseline variables and length of hospital stay as predictors of fear of falling.

	Univariable			Multivariable
Variable	OR (95% CI) Afraid of falling	P	Area under ROC curve (95% CI)	OR (95% CI) Afraid of falling	P
Age	1.03 (1.02–1.05)	0.0001	0.6 (0.55–0.66)
Sex
Female	2.42 (1.66–3.52)	<0.0001	0.61 (0.56–0.65)	2.25 (1.46–3.46)	0.0002
Male (Ref.)	1
NIHSS	1.1 (1.04–1.17)	0.0009	0.58 (0.53–0.63)
Length of hospital stay (days)	1.06 (1.04–1.08)	<0.0001	0.67 (0.62–0.72)
Type of stroke
Ischemic (Ref.)	1
Hemorrhagic	0.75 (0.35–1.6)	0.46	0.51 (0.49–0.54)
Subarachnoid	–	–	–
Stroke location
Right (Ref.)	1
Left	0.69 (0.47–1.01)	0.055
Bilateral	0.64 (0.17–2.44)	0.51	0.55 (0.50–0.59)
SwePASS total score
≤24	13.33 (7.47–23.8)	<0.0001		9.41 (5.13–17.25)	<0.0001
25–30	3.86 (2.44–6.12)	<0.0001		2.29 (1.36–3.83)	0.0017
≥31 (Ref.)	1		0.74 (0.70–0.78)	1
Using a walking aid	4.65 (3.11–6.95)	<0.0001	0.68 (0.64–0.72)	3.40 (2.12–5.43)	<0.0001
Hypertension	1.22 (0.81–1.85)	0.33	0.52 (0.48–0.56)
Atrial fibrillation	1.56 (1.03–2.37)	0.038	0.54 (0.50–0.58)
Heart failure	1.19 (0.65–2.15)	0.57	0.51 (0.48–0.54)
Ischemic heart disease	1.15 (0.74–1.8)	0.54	0.51 (0.47–0.55)
Diabetes mellitus	1.07 (0.68–1.67)	0.78	0.51 (0.47–0.54)
Hyperlipidaemia	0.67 (0.44–1.03)	0.067	0.54 (0.50–0.58)
Smoking
Never smoker (Ref.)	1
Previous smoker	0.74 (0.46–1.17)	0.20
Current smoker	0.70 (0.43–1.13)	0.15	0.54 (0.49–0.59)
SGPALS
1 Physically inactive (Ref.)	1
2 Some light PA	0.34 (0.23–0.50)	<0.0001
3 Regular PA/training	0.31 (0.12–0.80)	0.015	0.63 (0.59–0.68)

Notes:

The SwePASS: The Swedish modified version of the Postural Assessment Scale for Stroke Patients; OR: odds ratio; CI: confidence interval; Area under ROC curve: Area under the receiver operating characteristic curve; SGPALS: Saltin-Grimby Physical Activity Level Scale; PA: physical activity.

Discussion

This study investigated factors associated with FoF during the acute phase after stroke. The results of the multivariable regression analysis demonstrated that FoF was associated with poor postural control, female sex and the use of a walking aid. To our knowledge, the current study is unique due to the large sample size and the fact that it is one of very few studies to investigate the association between FoF and personal baseline factors in acute stroke.

FoF is very relevant for patients’ activation, participation and quality of life. For optimal rehabilitation, to adjust plans and approaches, we believe that it is important to pay attention to any FoF, as it could constitute a mobility barrier for the patient, an obstacle to challenging the daily training/exercise.

The results of the current study are consistent with a previous study of people with chronic stroke¹¹ (mean 53 months since stroke) that demonstrated that those reporting FoF had significantly worse balance than those without FoF. Poor postural control has also shown a strong association with falling.²⁹ However, in terms of falls, factors other than FoF seem to be of importance.

As previously presented, FoF is common among patients with both acute^27,29 and sub-acute stroke.¹⁴ Interestingly, our results show that women seem to be more afraid of falling, while according to the FallsGOT, men are more likely to fall.²⁹ The fact that the men did not report FoF as much as the women may be related to society’s expectations of how men and women are expected to behave, although this conclusion is speculative.

A majority (72%) of the patients in our study who used a walking aid reported FoF. The walking aid may contribute to an increased activity level if it enhances the patient’s sense of security. However, previous research on poststroke falls³ has reported that many falls (58%) in a stroke rehabilitation setting occurred because the patient was not following the rehabilitation team’s recommendations regarding the walking aid and/or the need for supervision. Among these patients, the majority were cognitively impaired.³

According to the International Classification of Functioning, Disability and Health, fear is a subjective experience that is classified and described as a body function and an emotional function.³⁸ As such, FoF in the acute phase after stroke might negatively affect rehabilitation,²⁷ and the experience of FoF in acute stroke may lead to a vicious circle if inactivity increases as a result of fear. Additionally, the consequences of inactivity could lead to decreased self-efficacy and greater FoF. Reduced self-efficacy in everyday activities could lead to avoidance of such activities. Avoidance behavior, and thus reduced activity level, might in turn affect the patient’s functioning, including postural control, muscle function, mobility and walking ability. Physiotherapists have an important task to perform in improving the strength of patients who are experiencing FoF and preventing the emergence of negative developments. The results of earlier research^{15–17,39,40} have shown that FoF among the elderly who have not experienced stroke contributes to decreased activity. In our study, in line with previous research among the elderly,^{15–17,39,40} older patients were more likely to report FoF. However, the association between age and FoF did not remain statistically significant in the multivariable regression analysis.

A stroke in the non-dominant hemisphere (the right hemisphere, in most people), may lead to reduced self-awareness, anosognosia, and neglect.⁴¹ Patients with a right-sided stroke can therefore be expected to report less FoF due to possible impairments in disease- and self-awareness. However, in our study, no difference based on stroke location was found. This is in contrast to a study of FoF²⁶ that found a negative correlation between BBS and self-perceived sense of balance in individuals with chronic stroke and left hemiplegia.²⁶

The strengths of the study are its consecutive enrollment and large population. The study fills a gap in the knowledge concerning FoF and associated factors in acute stroke. Since the majority of the patients had low NIHSS scores, the results primarily describe FoF in patients with limited stroke severity. Therefore, the results cannot be generalized to patient populations other than those with mild to moderate stroke. Another limitation is the possibility that cultural factors affect FoF; therefore, the results may not be globally generalizable. Moreover, there was a selection bias because patients being considered for thrombolysis or thrombectomy were not admitted to the stroke unit at SU/Östra. Since the ambition was to examine any FoF in a fast and easy way, the single question “Are you afraid of falling?” was chosen. However, it might have been prone for bias. A standardized questionnaire/scale might have been more feasible. The current study, based on one structured closed question, is a first step in describing FoF and its associations in the acute phase of stroke. However, to gain deeper knowledge, the authors suggest that future studies use a qualitative approach. Also, the association between FoF and future falls needs further investigation. Asking about FoF prior to stroke onset, in an epidemiological study, might also be of interest.

Conclusion

Our findings provide valuable insight for those involved in stroke rehabilitation during the acute phase after stroke. FoF is associated with poor postural control, female sex and the use of a walking aid.

Clinical trial registration

ClinicalTrials.gov, Identifier: NCT02264470.

Declaration of conflicting interests

The authors declare no potential conflicts of interests with respect to the research, authorship and/or publication of this article.

Funding

This work was supported by the Healthcare sub-committee, Region Västra Götaland, the Healthcare Board Region Västra Götaland, Sahlgrenska University Hospital Foundation, Felix Neubergh’s Foundation, Stroke Centrum West, Greta and Einar Asker’s Foundation, Renée Eander’s Foundation and the Foundation Hjalmar Svensson research fund.

Acknowledgements

We would like to thank Aldina Pivodic at Statistiska Konsultgruppen, for statistical consultation.