10   SA JournAl of PhySiotherAPy 2013 Vol 69 no 2

Research

Article

FUNctiONal OUtcOMes OF 
stROKe PatieNts aDMitteD tO a 

teRtiaRy hOsPital iN the 
WesteRN caPe, sOUth aFRica

Correspondence Author:
Ishita Parekh
103 Jabral Court,
22 Main Road, Rosebank 7700
Cape Town
Republic of South Africa
Email: iparekh87@gmail.com

AbstrACt:  Individuals with stroke often experience functional limitations 
such as the inability to walk independently. There are several factors that are 
associated with functional outcome post stroke. The aim of the study was therefore 
to determine functional outcomes and factors influencing functional outcomes of 
stroke patients admitted to a South African tertiary hospital. A longitudinal design 
was used to obtain the data. A convenient sampling method was used to recruit 
100 participants. Data was collected on admission, discharge and two months post 
stroke, using the national institute of health Stroke Scale, barthel index and a 
socio-demographic and medical profile data sheet. The data was analyzed using 
SPSS version 18 and SAS version 9.  Tobit analysis was used to determine the factors influencing functional outcomes at discharge 
and two months post stroke. The necessary ethical clearance was obtained prior to commencement of the study. Stair climbing, 
mobility and transfers were the functional activities that the majority of the participants could not perform independently. Age and 
severity of stroke significantly influenced the functional outcomes at discharge and two months post stroke. factors such as age 
and stroke severity should therefore be considered when planning rehabilitation interventions for stroke patients in this setting. 

Key words:  STRoke, funcTionAl ouTcomeS, SeveRiTy of STRoke, SouTh AfRicA

Parekh i (Msc)1

Rhoda a (PhD)1

 1 University of the Western Cape, 
  South Africa.

experienced a stroke have limitations 
in performing self­care, household and 
community activities (Rhoda 2011).

Various studies identified different 
factors associated with outcome post 
stroke (Meijer et al 2003; Gialanella 
et al 2012). Disability on admission, 
swallowing problems, age, severe 
paresis and urinary incontinence were 
found to be predictors of functional 
outcome post stroke (Meijer et al 2003).  
Ability to perform basic activities 
of daily living is also an important 
predictor of stroke outcome (Gialanella 
et al 2012). Furthermore, factors relating 
to the structure of rehabilitation have 
been found to affect the outcome post 
stroke. Strong evidence exists for the 
rehabilitation of patients with stroke 
in specialized stroke units. Organized 
inpatient stroke units are effective in 
improving short­term survival, func­
tional ability, independence and the 
need for institutionalization (Stroke  
Unit trialists’ Collaboration 2007). In 

bacKGROUND
Stroke is a common global health care 
problem that is serious and disabling 
(Warlow et al 2008). If viewed within the 
International Classification of Functional 
Disability and Health, the impact of 
stroke can be characterised by a number 
of impairments, activity limitations 
and participations restric tions that 
are experienced post stroke. The most 
common impairments identified are 
motor and sensory function, abnormal 
muscle tone, cognitive and perceptual 
limitations, speech impairments, depres­
sion, urinary incontinence and dysphagia 
(Lawrence et al 2001). Patients who have 

South Africa, the stroke unit model of 
care has not been widely implemented 
despite robust evidence of its efficacy 
(Bryer et al 2010). Further research is 
needed regarding the management of 
stroke patients in stroke units in South 
Africa. In the recent years information 
relating to outcomes of stroke patients 
admitted to other levels of healthcare in 
the Western Cape have been published 
(Rhoda et al 2011; Rouillard et al 2012), 
but no published literature is available 
regarding the outcomes of patients dis­
charged from a tertiary hospital which 
has a stroke unit in the Western Cape. 
Milinaviciene et al (2007) stated that 
for effective utilization of rehabilitation 
resources, identifying various factors 
affecting functional outcome is impor­
tant. The aim of the current study was 
therefore to determine the functional 
outcomes and factors affecting these 
outcomes of stroke patients admitted 
to a tertiary acute care hospital in the 
Western Cape, South Africa.



11   SA JournAl of PhySiotherAPy 2013 Vol 69 no 2 

MethODOlOGy
The study was conducted on 100 stroke 
patients consecutively admitted at 
Groote Schuur Tertiary hospital, Cape 
Town for a 7­month period that occurred 
from 1 October 2010 to 30 April 2011. 
A descriptive, longitudinal, quantitative 
design was used. A convenient non­
probability sampling method was used. 
The inclusion criteria were first ever 
occurrence of stroke, onset ≤4 weeks 
prior to admission and provision of 
written informed consent either by 
the patient or the caregiver. Patients 
were excluded from the study if they 
had additional neurological conditions 
such as spinal cord injury, head injury, 
dementia or history of seizures prior to 
stroke. Patients with speech problems 
were included in the study. 

A self­developed, socio­demographic 
and medical profile data sheet was used 
to capture the socio­demographic and 
medical profile data from the medical 
records of the patients. Data collected 
on admission included age, gender, 
level of education, employment status, 
documented risk factors for stroke, 
type of stroke, side of lesion and side of 
impairment. The variables captured from 
participants’ folders on discharge were 
frequency and type of rehabilitation 
and length of hospital stay. The face 
and content validity of the data sheet 
was determined by it being reviewed by 
experts in the field of stroke manage­
ment. The National Institutes of Health 
Stroke Scale (NIHSS) (Brott et al 1989) 
was used to determine the severity of 
stroke on admission and discharge. The 
scale is completed by direct observation 
as opposed to self­report. The NIHSS 
includes the following 15 items; 
cognitive impairment, gaze deviation, 
hemianopia, facial palsy, weakness of 
lower and upper limbs, ataxia of lower 
and upper limbs, sensory loss, dysarthria, 
aphasia and inattention. The scale is 
scored from 0–42, with a higher score 
indicating a higher level of neurological 
impairment or stroke severity. The 
intraclass correlation coefficient for the 
NIHSS was documented as between 
0.93 to 0.95 (Godstein and Samsa 1997). 
The scale was found to be valid and 
shows concurrent validity with the 
Barthel Index (BI). Good agreement has 

analyzed using SPSS version 19 and 
Statistical Analysis System (SAS) ver­
sion 9. Descriptive statistics such as 
mean, median, mode and standard 
deviation were calculated using SPSS. 
Due to the type and distribution of the 
data Spearman’s correlation was used 
to determine the correlation between 
the outcome variable which was the 
Barthel score at discharge and two 
months post stroke and each potential 
predictor variable (age, frequency of 
rehabilitation, severity at admission and 
functional level at admission). Tobit 
analysis was used to determine factors 
influencing the outcome variable at 
discharge and 2 months post stroke. The 
Tobit model also called the censored 
regression model (Yang and Douglas 
2010) was used to address the ceiling 
and floor effects noted in the out come. 
The significance level for inclusion in  
the final regression was p< 0.0001. 
The data of the patients who died, at 
discharge (n=6) and 2 months (n=6) 
were included in the analysis and BI 
score, as zero was used.

ethics
Ethical clearance to conduct the study 
was obtained from the Senate Higher 
Degrees Committee of the University of 
the Western Cape and the Senate research 
grants and study leave committee of 
UWC. Permission was also obtained 
from the Medical Superintendent of  
Groote Schuur Hospital and the Chair­
person of Health Science Faculty 
Research Ethics Committee at University  
of Cape Town. The aims, objectives 
and nature of the study were explained 
to all the participants. If they agreed 
to participate in the study then written 
informed consent was obtained from 
them or a primary caregiver. The par­
ticipants were informed of their right to 
withdraw and that confidentiality and 
anonymity would be maintained during 
the collection, analysis and publication 
of the data.

ResUlts
socio-demographic and medical 
fac tors
A total of 100 participants were recruited 
to be part of the study. The mean age of 
the participants was 59.83(SD=15.71) 

been documented between correlation 
coefficients when comparing the NIHSS 
with the BI, the Rankin Scale, and the 
Glasgow Outcome (Lyden et al 1999). 

The Barthel Index (BI) (Mahoney 
and Barthel 1965) was used to capture 
data relating to the functional ability of 
the participants. The aim of the scale is 
to measure the level of independence 
in the basic activities of daily living. 
The scores range from 0–100 with 
higher scores indicated higher level of 
functioning (Finch et al 2002). Infor­
mation needed to complete the BI could 
be obtained from the most reliable 
source, which could be the health care 
worker or the carer (Finch et al 2002). 
In the current study the researcher 
consulted with the nursing staff and the 
patients caregiver when completing the 
BI. An excellent inter­rater reliability of 
r=0.489 (p<0.001) has been found for 
this tool (Rollnik 2011). The reliability 
of telephonic interview was also found 
excellent (95% CI, 0.85­0.94) (Pietra et 
al 2011).

The researcher visited the medical 
ward and stroke unit on a daily basis to 
identify patients who could be included 
into the study. Patients who were 
eligible to be included were invited to 
participate in the study with the aim 
of the study being explained to them. 
Written informed consent was obtained 
from relevant patients or proxies when 
the patient was unable to respond. Data 
were collected on admission, discharge 
and 2 months post stroke. After com­
pleting the socio­demographic and 
medical profile data sheet, the researcher 
completed the NIHSS and BI as above. 
These instruments were completed in 
privacy at a time that was convenient 
for the patient. The same instruments 
were completed with each participant 
at discharge with a window period of 
two days. At 2 months post discharge, 
telephonic interviews for the BI were 
conducted. The caregivers of patients 
who were unable to be interviewed via 
the telephone were interviewed as proxy 
respondents.

Data aNalysis
The data was captured using the 
Statistical Package for the Social 
Sciences (SPSS) version 19 while 



12   SA JournAl of PhySiotherAPy 2013 Vol 69 no 2

years with 51% females and 49% males. 
Hypertension was found to be the most 
common risk factors 79%, with smoking 
being the second most common risk 
factor 49% among the participants. 
Ischemic stroke was found in 89% 
while hemorrhagic stroke was in 10%. 
Six percent died before discharge and 
6% after discharge but before 2 months 
follow­up assessment period.

Rehabilitation process factors
The number of patients that received 
Physiotherapy (75%) and Occupational 
Therapy (79%) was nearly similar. The 
mean length of stay of patients was 
10.4(SD=5.80) days. More than half 
of the patients 53.1% were referred for 
follow­up to inpatient reha bilitation 
centers and 46.8% were referred for 
follow­up to outpatient rehabilitation.

severity of stroke
The mean of NIHSS was 12.49 and 8.76 
on admission and discharge respecti­
vely. Missing data of one participant 
and demise of six participants reduced 
the total number of patients as ninety­
three for the discharge NIHSS findings. 
On admission 21% of the patients 
had a severe stroke, which reduced at 
discharge and remained severe for 9.6% 
of the participants excluding those who 
died. The number of participants with 
minor stroke increased from 22% on 
admission to 30.2% at discharge. The 
majority of participants experienced a 
moderate stroke with 52% on admission 
and 44% at discharge.

Functional outcome
The mean and standard deviation (SD) 
of the BI scores was greater at 2 months 
55(40.3) than on discharge 43(35) and 
admission 28(29). 

Figure 1 illustrates the BI scores of 
the participants at the three assessment 
points; admission, discharge and 2 
months post stroke. The majority (77%) 
of participants were totally dependent  
on others to perform functional activi­
ties on admission. Functional outcome 
increased over time but 43.9% of 
participants were still dependent at 2 
months post stroke.

Table 1 presents the number of 
participants who were independent on 

P
e
rc

e
n

ta
g

e
 o

f 
p

a
rt

ic
ip

a
n

ts

Figure 1: Functional outcome as measured by bi at different assessment periods

table 1: Descriptive statistics of bi

Individual items
Admission (n=100)
Number (%)

Discharge (n=99)
Number (%)

2 months (n=66)
Number (%)

Feeding 61(61) 77(77.7) 51(77.2)

Bathing 57(57) 69(69.6) 48(72.7)

Grooming 53(53) 65(65.6) 46(69.6)

Dressing 42(42) 57(57.5) 43(65.1)

Bowel 33(33) 48(48.4) 41(62.1)

Bladder 25(25) 42(42.4) 39(59.0)

Toilet 23(23) 39(39.3) 38(57.5)

Transfer 13(13) 29(29.2) 35(53.0)

Mobility 2(2) 9(9.0) 26(39.3)

Stairs 0 1(1.0) 4(6.0)



13   SA JournAl of PhySiotherAPy 2013 Vol 69 no 2 

different items of the BI at different 
assessment periods. The activities that 
the majority of participants needed 
assistance with at all three assessment 
periods, was stair climbing followed by 
mobility and transfers.

correlation between outcome 
va riable and potential variable
The results of the Spearman’s correlation 
indicated that the total number 
of rehabilitation sessions was not 
significantly correlated with functional 
outcome at discharge (p=0.0842). At 
discharge, age (p<0.0001), severity of  
stroke on admission (p<0.0001), 
and functional levels on admission 
(p<0.0001), correlated significantly with 
functional outcome while at 2 months 
post stroke, age and severity of stroke on 
discharge correlated significantly with 
functional outcome. These variables 
were therefore included appropriately in 
the regression analysis to determine the 
factors influencing functional outcome 
at discharge and 2 months post stroke.

Factors influencing functional out-
come at discharge
The factors that were identified as those 
influencing functional outcome were 
age, stroke severity and functional level 
on admission. The analysis of maximum 
likelihood parameter estimate showed 
that the coefficient for age was ­0.50.  
For each year increase in age, the 
estimated mean would decrease by 
0.50. For every ten year increase in age, 
the discharge functional score would 
decrease by 5­points. The coefficient 
for functional level at admission was 
0.55. For 1­point increase in admission 
functional score, the discharge func­
tional score would increase by 0.5. The 
coefficient for severity at admission 
was 1.75. A 1­point increase in severity 
would correspond to 1.75­point decrease 
in discharge functional score.

Factors influencing functional out-
come at 2 months
The analysis of maximum likelihood 
parameter estimated that age (p=0.0143) 
and severity at admission (p<0.0001) 
were both significant together for 
functional outcome at 2 months. The 
coefficient for age was ­0.61. Each one 

year increase in age would decrease 
the functional score by 0.61. Each ten 
year increase in age would decrease it 
by 6­points. The coefficient for severity 
at admission was ­3.50 meaning each 
1­point increase in severity at admission 
corresponded to an estimated 3.50 
decrease in functional outcome at two 
months.

DiscUssiON

Participant characteristics
A total of 100 participants were recruited 
to be part of the study. The mean age 
of the participants, 59.83(SD=15.71) 
years, was younger than stroke patients 
admitted to inpatient facilities in 
developed countries, which was reported 
as 75 years (Lee et al 2011).  The mean 
age of the participants were however 
similar to stroke patients admitted to 
inpatient rehabilitation centers in the 
Western Cape. Rouillard (2006) and 
Bryer (2009) reported the age of stroke 
patients admitted to inpatient facilities 
in the Western Cape as 51.3 years and 
59 years respectively. The patients were 
however slightly younger than the stroke 
patients managed at Community Health 
Centres (CHC) in the Western Cape 
where a mean age of 61(SD=10.5) years 
was reported (Rhoda et al 2011). It is not 
clear why the mean age of the patients 
is much lower in sub­Saharan African 
countries than other countries. From the 
results of these local studies it appears 
that stroke patients being admitted for 
inpatient management are younger than 
those managed at outpatient facilities. 
The limited resources for rehabilitation 
in developing countries could be the 
reason for this as younger patients are 
admitted for intense rehabilitation as 
they have better outcomes than older 
ones (Varona 2011).

As found in previous studies (Rhoda et 
al 2003; Rouillard et al 2012) conducted 
in the Western Cape, hypertension was 
the most prevalent risk factor for stroke. 
It was found that the knowledge of stroke 
risk factors was inadequate among the 
South African population (Bogoshi et al 
2003). This lack of knowledge about the 
risk of stroke could be resulting in stroke 
being a cause of morbidity and mortality 
in this country.

Functional outcome of stroke on 
admission, discharge and 2 months
The functional outcomes of the indi­
viduals were determined using the BI. 
The participants admitted to the tertiary 
hospital needed assistance with mobi­
lity, stair climbing and transfers. With 
exception of stair climbing these were 
different to the activities that patients 
in the previous studies conducted in the 
Western Cape needed assistance with. 
Participants in the studies conducted by 
Rhoda et al (2011) and Rouillard et al 
(2012) were dependent in activities such 
as feeding, washing and stair climbing. 
In the present study the majority of 
participants were dependent in daily 
activities, scoring between 0 and 55 on 
BI at the time of admission (Granger 
et al 1979). It should however be noted 
that in the acute stage, the dependent 
state does not always mean inability 
to perform activities of daily living it 
could be linked medical directive. At 
discharge the majority of participants 
were still however dependent. The 
reason for this could be the short 
length of stay (mean=10.4 days) which 
patients admitted to tertiary institutions 
experience. At 2 months, a significant 
improvement was however noted for 
individual items on BI scores. The 
majority (43.9%) of participants still 
however needed maximum assistance. 
There is also a concern that early 
benefits achieved by acute stroke unit 
rehabilitation could be eroded by a lack 
of adequate rehabilitation after discharge 
(De Villers et al 2009). The study of 
Wang et al (2011) suggests that earlier 
transfer to an inpatient rehabilitation 
hospital leads to improved functional 
outcome. However, certain factors such 
as age, initial medical condition and 
functional status contribute to functional 
improvement.

Factors affecting functional out-
comes at discharge and 2 months 
post stroke
Increasing age was found to be 
significantly associated with poor func­
tional outcome (p<0.0001) at discharge 
and two months. This finding could 
be as a result of additional disabilities  
and comorbidities, which are associated 
with increased age (Nichols­Larsen et 



14   SA JournAl of PhySiotherAPy 2013 Vol 69 no 2

al 2005). As confirmed by the findings 
of this study severity of stroke was an 
important factor, which could be used to 
determine the outcome post stroke. This 
is an important finding in the countries 
like South Africa where the resources 
for stroke rehabilitation is limited which 
could consider stroke severity as an 
admission criteria. It should however be 
noted that if structured interdisciplinary 
stroke units are available these patients 
with severe stroke have improved 
outcomes.

Functional level on admission pre­
dicted functional level at discharge and 2 
months post stroke. This finding supports 
that of a previous study conducted 
by Tilling et al (2001). The level of 
functional ability at admission could 
therefore be used to predict functional 
level at discharge and be used to assist 
with discharge planning.

cONclUsiON
It can therefore be concluded from the 
results of the study that age, severity of 
stroke and functional level at admission 
are the factors that affect functional 
outcome at discharge and 2 months post 
stroke of patients admitted to a tertiary 
hospital in the Western Cape, South 
Africa. Therefore, the rehabilitation 
staff should consider these factors while 
planning rehabilitation interventions 
and referral to rehabilitation facilities 
for patients admitted to this setting. In 
addition rehabilitation of elderly patients 
could focus more on task oriented 
activities and compensation rather than 
on recovery of neurological status.

ReFeReNces

Bogoshi GM, Stewart A, Hale L, Fritz V 2003 
Knowledge of stroke risk factors amongst black 
diabetic, hypertensive and stroke patients. South 
African Journal of Physiotherapy 59(4):25­30

Brott HP, Adams CP, Olinger JR, Marler WG, 
Barsan J, Biller J, Spilker R, Holleran R, Eberle 
V, Hertzberg M, Rorick CJ, Moomaw MW 1989 
Measurements of acute cerebral infarction: a clini­
cal examination scale. stroke, 20:964­970

Bryer A 2009 The need for a community­based 
model for stroke care in South Africa. South 
African Medical Journal Cape Town 99(8)

Bryer A, Connor M, Haug P, Cheyip B, Staub H, 
Tipping B, Duim W, Pinkney­Atkinson V 2010 
South African guideline for management of 
ischaemic stroke and transient ischaemic attack: 
a guideline from the South African Stroke Society 
(SASS) and the SASS Writing Committee. South 
African Medical Journal 100:747­78

De Villiers L, Kalula S, Burch V 2009 Does 
multidisciplinary stroke care improve outcome 
in a secondary level hospital in South Africa? 
International Journal of Stroke 2:89­93

Finch E, Brooks D, Stratford PW, Mayo NE 2002. 
Physical rehabilitation outcome measures: A guide 
to enhanced clinical decision­making Baltimore 
(MD): Lippincott Williams & Wilkins

Gialanella B, Santoro R, Ferlucci C 2012 Predicting 
the outcome after stroke: the role of basic activities 
of daily living. European Journal of Physical and 
Rehabilitation Medicine (ahead of print) 

Goldstein LB, Samsa GP 1997 Reliability of 
the National Institutes of Health Stroke Scale. 
Extension to non­neurologists in the context of a 
clinical trial Stroke 28(2):307­310  

Granger C, Dewis L, Peters N, Sherwood C, Barret J 
 1979 Stroke Rehabilitation: analysis of repeated 
Barthel Index Measures. Archives of Physical 
Medicine and Rehabilitation 60:14­17

Lawrence E, Coshall C, Dundas R, Stewart J, 
Rudd AG, Howard R, Wolf CD 2001  Estimates 
of the prevalence of acute stroke impairments 
and disability in a multiethnic population. Stroke 
32(6):1279­84

Lyden P, Mei L, Jackson C, Marler J, Kothari R, 
Brott T 1999 Underlying structure of the National 
Institutes of Health Stroke Scale: Results of a fac­
tor analysis. Stroke, 30:2347­2354

Lee S, Shafe A, Cowie M 2011 UK Stroke 
incidence, mortality and cardiovascular risk 
management 1999­2008: time­trend analysis from 
the General Practice Research Database. British 
Medical Journal Open 1(2):e000269

Mahoney FI, Barthel DW 1965 Functional evalu­
ation: the Barthel Index. Maryland state medical 
journal Feb;14:61­65  

Meijer R, Ihnenfeldt DS, de Groot IJ, van Limbeek 
J, Vermeulen M, de Haan RJ 2003 Prognostic 
factors for ambulation and activities of daily living 
in the subacute phase after stroke. A systematic 
review of the literature. Clinical Rehabilitation 
Mar;17(2):119­129

Milinaviciene E, Rastenyte D, Krisciūnas A 2007 
Factors influencing outcomes in patients with 
stroke. Medicina (Kaunas) 43(4):269­77

Nichols­Larsen DS, Clark PC, Zeringue A, 
Greenspan A, Blanton S 2005 Factors influencing 
stroke survivors’ quality of life during subacute 
recovery. Stroke 36(7):1480­1484

Pietra GLD, Katia S, Elodie O, Monica R, 
Francesco M, Maurizio AL 2011 Validity and 
Reliability of the Barthel Index Administered by 
Telephone. Stroke 42:2077­2079

Rhoda A, Hendry J 2003 Profile of stroke patients 
treated at a Community­Based rehabilitation Centre  
in a Cape Town Health District. South African 
Journal of Physiotherapy  59(4):20­24

Rhoda A, Mpofu R, De Weerdt W 2011 Activity 
limitations of patients with stroke attending  
out­patient facilities in the Western Cape, South 
Africa. South African Journal of Physiotherapy 
67(2):16­22

Rollnik JD 2011 The early rehabilitation Barthel 
Index (ERBI). Die Rehabilitation 50(6):408­411

Rouillard S 2006 Pattern of recovery and outcome 
after stroke in patients accessing a Western Cape 
rehabilitation facility. Cape Town 

Rouillard S, Willy De Weerdt W, De Wit L, 
Jelsma J 2012 Functioning at 6 months post stroke 
following discharge from inpatient rehabilitation. 
South African Medical Journal 102(6):545­48

Stroke Unit Trialists’ Collaboration 2007. Orga­
nised inpatient (stroke unit) care for stroke. 
Cochrane Database of Systematic Reviews, Issue 
4. Art. No.: CD000197

Varona J 2011 Long­term prognosis of ischemic 
stroke in young adults. Stroke Research and 
Treatment. Vol. 2011,Article ID 879817, 5 pages, 
doi: 10.4061/2011/879817

Warlow CP, van Gijn J, Dennis MS, Wardlaw J, 
Bamford J, Hankey G, Sandercock P, Rinkel P, 
Langhorne P, Sudlow C, Rothwell P 2008 Stroke: 
practical management (3rd edn.) Blackwell 
Publishing, Oxford

Wang H, Camicia M, Terdiman J, Hung Y, Sandel 
E 2011 Time to Inpatient Rehabilitation Hospital 
Admission and Functional Outcomes of Stroke 
Patients. Physical Medicine and Rehabilitation 
3:296­304

Yang Y, Simpson DG 2010 Conditional decom­
position diagnostics for regression analysis of 
zero­inflated and left­censored data. Statistical 
Methods in Medical Research 0(0):1­16