Clinical Rehabilitation 1999; 13: 268–272

Post-stroke depression and functional outcome:

a cohort study investigating the influence of
depression on functional recovery from stroke
FB van de Weg Rehabilitation Centre, Amsterdam, DJ Kuik Department of Epidemiology and Biostatistics, Vrije
Universiteit Amsterdam and GJ Lankhorst Department of Rehabilitation Medicine, University Hospital Vrije Universiteit,
Amsterdam, The Netherlands

Received 30th October 1997; returned for revisions 27th April 1998; revised manuscript accepted 7th October 1998.

Objective : To investigate the influence of depression on functional recovery

after stroke.
Design : Multicentre cohort study of 85 patients admitted for clinical
rehabilitation. A two-stage case-finding procedure was used to identify
patients with depression. For the control group, consecutive nondepressed
stroke patients were enrolled. Patients were interviewed at 3–6 weeks and
six months after stroke onset.
Setting : Three rehabilitation centres in the vicinity of Amsterdam.
Main outcome measures : Functional outcome was determined by the
Functional Independence Measure (FIM) and the Rehabilitation Activities
Profile (RAP).
Results : The prevalence of depression (35%) was comparable with the
findings of earlier studies in other settings. Patients classified as depressed
according to DSM III R criteria (American Psychiatric Association Diagnostic
and statistical manual of mental disorders) had a significantly lower functional
score, both at onset and after follow-up (FIM and RAP). There was, however,
no significant difference in functional improvement between the depressed
and the nondepressed group. Mean functional improvement in the six patients
treated with antidepressants was 30% better than in the untreated
(depressed) patients; numbers were too small for the results to attain
statistical significance. Subset analysis showed a significantly higher outcome
for nondepressed patients for the FIM subitems personal care and transfers.
However, functional improvement was not significantly different for any of the
subitems in depressed versus nondepressed patients.
Conclusion : Stroke patients with depression have significantly lower
functional scores both at onset and after six months. Our results suggest
under-recognition of post-stroke depression and a possible beneficial effect of
antidepressant medication in depressed stroke patients. Further studies are
required to determine the effect of antidepressants.

Address for correspondence: FB van de Weg, Rehabilitation

Centre Amsterdam, Overtoom 283, 1054 HW Amsterdam,
The Netherlands.
© Arnold 1999 0269–2155(99)CR251OA

Introduction
In the US approximately 600 000 people a year
sustain a stroke; in 1986 there were over 2 mil-
lion stroke survivors.1 Recent studies have high-
lighted the issue of post-stroke depression (PSD)
and reported prevalencies vary from 25% to
79%.1 Common stroke sequelae, such as aphasia
or a flattened affect from frontal lesions make it
difficult to establish the true prevalence of PSD.
Another possible reason for the current variance
in findings with regard to the incidence of PSD
is the methodological heterogeneity of the vari-
ous studies. However, even the most conservative
estimate of the prevalence of PSD must amount
to approximately 500 000 in the US alone. This
figure emphasizes the fact that PSD is a major
problem which deserves full medical attention, in
particular because of the potential negative
impact of depression on patient participation in
the rehabilitation process and the associated
rehabilitation outcome.
Parikh et al.2 followed a cohort of 63 stroke
patients; the 25 depressed patients were signifi-
cantly more impaired at two-year follow-up than
the nondepressed group in both physical activi-
ties and language functions. Over 25% of the
depressed patients were aphasic, making it diffi-
cult to determine depression from a self-rating
scale. External validity was not optimal since the
study population was comprised predominantly
of lower socioeconomic class black patients. In a
cohort of 49 hospital patients, of whom 20 were
depressed (according to the American Psychiatric
Association Diagnostic and statistical manual of
mental disorders criteria, DSM III-R) at two
months after stroke, Morris et al.3 found that
depressed and nondepressed groups showed
equal improvement over baseline in activities
of daily living ADL (Barthel Index). Similar
findings were reported by Sinyor et al.4; 64
patients were examined on admission and at dis-
charge. The follow-up period was, however, not
indicated. Other studies found a significantly
lower functional score on admission in the PSD
group, yet no significant difference at follow-
up.5,6
Schubert7 examined 15 patients admitted for
rehabilitation. The change in Barthel Index (BI)
assessed in relation to depression was hampered
Stroke, depression and functional outcome 269
by the ‘ceiling effect’ of the BI, since initial scores
for nondepressed patients were already high and
thus unable to increase much more. Moreover,
follow-up was only four weeks after the initial
examination. Diamond et al.8 found that patients
in a geriatric rehabilitation unit who were classi-
fied as depressed at discharge had lower Func-
tional Independence Measure (FIM) scores on
admission and at discharge.
In summary, results from studies on functional
outcome in PSD are highly inconclusive. Inter-
pretation of the studies is complicated by the
methodological problems resulting from bias in
selection, sampling and/or confounding. We are
aware of only two studies7,8 examining depression
in the rehabilitation setting. We believe that it is
important to try to identify PSD, a common
sequela of stroke possibly sensitive to treatment.
There are encouraging signs that PSD responds
well to tricyclic antidepressant medication.1 How-
ever, further treatment studies are required
before the possible benefit of drugs can be defi-
nitely established. The aim of this study was to
investigate whether a depressive mood adversely
affects functional outcome in stroke patients six
months after stroke onset in a group of patients
referred for inpatient rehabilitation.
Method
Stroke was diagnosed by a neurologist, according
to the World Health Organization (WHO) crite-
ria.9 We performed a cohort study including all
stroke patients consecutively admitted to three
rehabilitation centres during the period Septem-
ber 1995 to June 1996 who fulfilled the following
criteria: (1) Referred for inpatient rehabilitation
between 3 and 6 weeks post stroke. (2) Able to
understand the interview questions and respond
reliably. This was established using the first
part of the Dutch translation of the Aphasia
screening test.10 (3) ‘Admission’ Functional
Independence Measure (FIM,11) score less than
110. Excluded were those who had: (1) a history
of manifest depression or treatment for depres-
sion (counselling, antidepressant medication,
electroconvulsive therapy) preceding stroke;
(2) concomitant conditions interfering with
assessment, e.g. dementia, severe osteoarthrosis;
270 FB van de Weg et al.
(3) psychotic depression or other psychiatric
disorder; (4) prior history of stroke.
The Dutch translation of the Geriatric Depres-
sion Scale (GDS,12) with a cut-off point of 8 was
used to screen for depression. The GDS consists
of 30 questions on a yes/no basis.
All patients who screened positively were then
subjected to the DSM III R criteria13 for a posi-
tive diagnosis of depression. All data were
acquired by one physician not related to the reha-
bilitation centre nor in any way involved in the
treatment of the patients. There was no contact
between the investigator and the primary care-
givers. Patients were assigned to one of two
groups: those with and those without a major
depressive disorder. Four patients who were
depressed at follow-up, but not at the time of the
first interview, were classified as nondepressed.
Age, sex, time since stroke onset, medication
and having a partner were recorded at the first
interview which took place 3–6 weeks after
stroke onset. Follow-up took place six months
post stroke. The FIM was used as primary out-
come parameter. The scaling characteristics of
the FIM and the slim likelihood of a ‘ceiling
effect’ occurring make it suitable for statistical
purposes in a study such as this.14 The communi-
cation, mobility and self-care domains of the
Rehabilitation Activities Profile (RAP) were
used as a secondary outcome parameter. The
RAP15 has good reliability, validity and respon-
siveness. A total of 12 activities are assessed on
a four-point scale (Table 1).
Statistical analysis was performed by means
of Fisher’s exact test and Student’s t-test for
unequal variances.
Results
The inclusion criteria were met by 87 patients, of
whom one patient died and one was lost to fol-
low-up. A total of 85 patients completed the
study: 42 men and 43 women with a mean age of
61.4 years (SD 11.4, range 27–81). Demographic
and clinical characteristics and mean functional
outcomes for the depressed and the control
group are shown in Table 2. Thirty patients were
classified as depressed according to the DSM III
R criteria: 16 men and 14 women indicating a
35% prevalence of a major depressive disorder in
this cohort. Only six patients were found to use
antidepressants at the first interview and three of
these patients were still depressed at follow-up.
Of the remaining 24, 52% were still depressed at
follow-up.
Table 3 shows the mean improvement in
functional outcome (FIM) for four patient
groups. A mean functional improvement of 30%
was found in the subgroup of depressed patients
treated with antidepressants (n = 6). Table 4
presents the mean outcomes for the six FIM sub-
sets for depressed and nondepressed patients.
Discussion
The prevalence of post stroke depression (35%)
was comparable with the findings of other
studies. We found no relationship between
sex/age/presence of a partner and depression.
Improvement in functional outcome was not
related to the presence of a partner, sex, or side
of hemiparesis. Mean RAP and FIM scores are
significantly lower for the depressed group both
at admission and at follow-up. Our results sup-
port the hypothesis of previous studies that
depression is associated with increased disability
in stroke patients. However, if outcome is repre-
sented in terms of functional improvement (i.e.
mean score at follow-up minus mean scores at the
Table 1 Rehabilitation Activities Profile (RAP)
communication, mobility and self-care domains
Activity
Expressing
Comprehending
Maintaining posture
Changing posture
Walking/using wheelchair
Climbing stairs
Using transport
Eating/drinking
Washing/grooming
Dressing
Undressing
Maintaining continence
Scores range from 0 to 3. Performs activity with:
0 no difficulty
1 some difficulty
2 much difficulty or some help
3 unable to perform activity.
 Stroke, depression and functional outcome 271

first interview) there is no significant difference
at all.
In general our results coincide with those of
Diamond et al.,8 who used the same dependent
and independent variables. However, they used
functional scores at discharge, thus introducing
length of inpatient stay as a possible confounder.
It is conceivable that the depressed patients in
our study and the study of Diamond et al.
improved relatively well because they were
offered the support and encouragement of an
intensive inpatient rehabilitation programme. In
any case, our results show that the potential for
functional improvement is no less in depressed
patients, although the eventual outcome is lower
in this group.

Table 2 Demographic and clinical characteristics and functional outcomes for depressed versus nondepressed
stroke patients

Characteristics Depressed Nondepressed

patients (n = 30) patients (n = 55) p-values

Male sex 16 (53.3%) 26 (46.7%) 0.82

Partner 11 (36.6%) 20 (36.8%) 0.81
Left hemiparesis 22 (73.3%) 28 (50.9%) 0.07
Geriatric depression scale positive 30 (100%) 24 (43.8%) –
Use of antidepressant 6 (20.0%) 0 (0%) 0.02

Mean age (SD) 62.5 (11.6) 60.5 (11.3) 0.44

Mean RAPa (admission) (SD) 16.8 (5.4) 13.7 (5.4) 0.02
Mean RAPa (follow-up) (SD) 10.5 (5.3) 7.0 (3.7) 0.03
Mean RAPa (follow-up – admission) (SD) 6.3 (4.3) 6.7 (4.2) 0.6
Mean FIMb (admission) (SD) 84.7 (11.8) 91.8 (12.6) 0.01
Mean FIMb (follow-up) (SD) 99.6 (10.9) 106.1 (7.3) 0.05
Mean FIMb (follow-up – admission) (SD) 14.9 (9.0) 14.4 (10.2) 0.8
a Rehabilitation Activities Profile (0–36).
b Functional Independence Measure (7–119).

Table 3 Mean improvement in functional outcome for four patient groups

Independent variable FIM (follow-up – admission) p-values

Partner/no partner 14.2 (10.2)/15.7 (8.4) 0.47

Left/right hemiparesis 14.3 (9.2)/15.2 (10.5) 0.66
Male/female 16.6 (10.5)/12.7 (8.6) 0.06
Antidepressants/no antidepressants 18.7 (12.1)/14.4 (9.5) 0.43

Values in parentheses are standard deviations.

Table 4 Mean outcomes of FIM subsets for depressed and nondepressed patients

Item FIM FIM at follow-up FIM (follow-up – admission)

Depressed Nondepressed p-value Depressed Nondepressed p-value

Personal care 26.9 (4.7) 29.8 (5.6) 0.004 5.1 (4.1) 5.2 (4.5) 0.9
Transfers 16.3 (3.6) 18.1 (1.7) 0.02 3.9 (3.7) 3.2 (3.8) 0.4
Locomotion 9.1 (3.5) 10.0 (3.0) 0.3 4.7 (3.6) 4.4 (3.4) 0.7
Communication 13.8 (0.6) 13.9 (0.4) 0.4 0.3 (0.6) 0.2 (0.5) 0.7
Social cognition 19.8 (1.6) 20.7 (0.7) 0.05 0.5 (1.4) 0.6 (1.4) 0.8
Sphincter control 13.6 (0.6) 13.5 (0.7) 0.8 0.4 (0.8) 0.7 (1.0) 0.2

Values in parentheses are standard deviations.

272 FB van de Weg et al.
A point worth noting is the possible under-
recognition of depression by the primary care-
givers, illustrated indirectly by the fact that only
6 out of 30 patients received antidepressant med-
ication even though physicians stated that, once
diagnosed, depression was usually treated with
antidepressants.
An improvement of 30% on the FIM score in
the treated group compared with the nontreated
group (Table 3) suggests efficacy. Although pos-
sibly clinically significant, numbers in this sub-
group are too small to attain statistical
significance. Regardless of whether or not med-
ication can ameliorate PSD, it is important to rec-
ognize depression and treat it accordingly: there
is a potential negative effect on the patient and
the rehabilitation process as a whole. Analysis of
FIM subsets and outcomes showed a significantly
higher outcome in the nondepressed group for
the items personal care and transfers. However,
functional improvement (FIM follow-up minus
admission) was not significantly different for any
of the six subitems in depressed versus non-
depressed groups. Our study confirms the find-
ings of other studies2,8 that a depressive mood is
associated with lower functional scores at onset
and at follow-up. In our opinion both the FIM
and the RAP are well-suited for this kind of mea-
surement; none of the patients had a maximum
follow-up score, thereby excluding the possibility
of a ‘ceiling effect’.
Some limitations of this study should be men-
tioned. No patients with aphasia were included,
thus limiting the external validity. We deliber-
ately chose not to include these patients.
Although they undoubtedly represent a sizeable
portion of the stroke population with depression,
there is at present no feasible and reliable
method available to screen them for depression.
A major drawback is that all patients were seen
at onset and at follow-up by the same investiga-
tor, which at least introduces the possibility of
observer bias. A possibility and a potential weak-
ness of this study is that some patients who
scored under the screening cut-off were
depressed. However, the likelihood of this occur-
ring is slim since a cut-off score of 8 is rather con-
servative, a fact reflected by the large number of
false positives after screening by the GDS
(43.8%, Table 2); in fact, we would recommend
a higher cut-off point in any future studies.
References
1 Gordon WA, Hibbard MR. Poststroke depression:
an examination of the literature. Arch Phys Med
Rehabil 1997; 78: 658–63.
2 Parikh RM, Robinson RG, Lipsey JR. The impact
of poststroke depression on recovery in ADL over a
2 year follow up. Arch Neurol 1990; 47: 785–89.
3 Morris PL, Raphael B, Robinson RG. Clinical
depression is associated with impaired recovery from
stroke. Med J Aust 1992; 157: 239–42.
4 Sinyor D, Amato P, Kaloupek D. Post stroke
depression: relationships to functional impairment,
coping strategies and rehabilitation outcome. Stroke
1986; 17: 1102–107.
5 Eastwood MR, Rifat SL, Nobbs N, Ruderman J.
Mood disorders following a cerebrovascular
accident. Br J Psychol 1989; 154: 195–200.
6 Feibel JH, Springer CJ. Depression and failure to
resume social activity after stroke. Arch Phys Med
Rehabil 1982; 63: 276.
7 Schubert DSP. Physical consequences of depression.
Gen Hosp Psychiatry 1992; 14: 69–76.
8 Diamond PT, Holroyd S, Macciochi SN, Felsenthal
G. Prevalence of depression and outcome on the
geriatric rehabilitation unit. Am J Phys Med Rehabil
1995; 74: 214–17.
9 WHO task force on stroke and other
cerebrovascular disease. Stroke 1989
recommendations on stroke prevention, diagnosis
and therapy. Stroke 1989; 20: 1407–31.
10 Rouma EJM. A.S.T.: Aphasia screening test
vertaald. T Logop Fon 1993; 1.
11 Ocszowski WJ. The FIM; its use to identify
rehabilitation needs in stroke survivors. Arch Phys
Med Rehabil 1993; 74: 1291–94.
12 Brink TL, Yesavage JA, Heersema PH, Adey M,
Rose TL. Screening tests for geriatric depression.
Clin Gerontol 1982; 1: 37–43.
13 American Psychiatric Association. Diagnostic and
statistical manual of mental disorders, third edition,
revised version (DSM III-R), Washington DC: APA,
1987.
14 Bunch WH, Dvonch VM. The ‘value’ of FIM scores.
Arch Phys Med Rehabil 1994; 73: 40–43.
15 van Bennekom CAM, Jelles F, Lankhorst GJ.
Rehabilitation Activities Profile: the ICIDH as a
framework for a problem-oriented assessment
method in rehabilitation medicine. Disabil Rehabil
1995; 17: 169–75.