Quality of Life of Stroke Survivors: A 1-Year Follow-Up Study

Methods 

Results 

During the study period, 392 patients with stroke were admitted to PWH. Of these, 303 (77%) were recruited and followed up at 3 months after stroke onset. The reasons for excluding the 89 patients were as follows: handicapped before stroke (n=24), death before 3 months (n=38), recurrence of stroke (n=14), terminal illness (n=5), non-Chinese ethnicity (n=1), unable to contact (n=4), and refusal to participate (n=3).

Of the 303 subjects, 269 (88.8%) subjects had ischemic strokes; 13 (4.3%) subjects had cardioembolic strokes; and 21 (6.9%) subjects had hemorrhagic strokes. The NIHSS score on hospital admission was less than 3 (indicating mild stroke) for 49 (16%) patients, between 3 and 8 (moderate stroke) for 174 (57%) patients, and more than 8 (severe stroke) in 80 (26%) patients. At 3 months, a total of 166 (56%) patients required personal assistance; sources of assistance were a spouse (n=46), other family members (n=30), formal caregivers (n=81), and others (n=9).

The baseline demographic and clinical characteristics of the 303 patients are shown in table 1. The mean age was 70.5 years and the ratio of men to women was about 1:1. About 1 in 5 patients lived in a nursing home. Twenty-six percent were recipients of comprehensive social security assistance, which was a social welfare payment for subsistence living of the poor in Hong Kong (US $300/mo). Twenty-four percent received disability allowance, which was a non–means-tested welfare payment for the disabled, amounting to US $120 to $300 per month. Twenty-six percent received old age assistance, which was a non–means-tested welfare payment for all persons 70 years or older, amounting to US $100 per month.

Table 1. Demographic and Clinical Characteristics of 303 Stroke Survivors at 3 Months

	Characteristic	Values
	Age (y)	70.5±11.6
	Sex
	Female	150(50)
	Male	153(50)
	Residence
	Old-age home	58(19)
	Own home	245(81)
	Marital status
	Married	184(61)
	Widowed/single/divorced	119(39)
	Education (y)
	<3	126(42)
	≥3	177(58)
	Welfare assistance
	Comprehensive social assistance	80(26)
	Disability allowance	72(24)
	Old-age allowance	79(26)
	None	72(24)
	Physical exercise per day
	≥1h	45(15)
	20–59min	119(39)
	<20min	139(46)
	Pain in affected side
	None	212(70)
	Mild	61(20)
	Moderate/severe	30(10)
	Dysphagia
	No	212(70)
	Mild	58(19)
	Moderate/severe	33(11)
	Diet
	Soft/puree	33(11)
	Tube feeding	23(8)
	Normal	247(81)

NOTE. Values are mean ± standard deviation or n (%).

Most patients performed less than 1 hour of physical exercise a day. About a third of patients reported mild or moderate-to-severe pain in the affected side, and about a third reported mild or moderate-to-severe dysphagia. Approximately one fifth could eat only soft food or required tube feeding.

At 3 months, 139 (46%) patients were independent in all basic ADLs, and achieved the maximum Barthel Index score of 20. In contrast, 73 (24%) patients were very dependent in basic ADLs and had a Barthel Index score of less than 15. Some 36% (94/263) of patients supplying reliable data had a GDS score of at least 8, which suggested clinically significant depression.7

HRQOL data at 3 months were available for 247 (82%) patients. Their median HRQOL score for the 4 domains ranged from 56% to 69%, being lowest in the physical and psychological domains (table 2). The reasons for missing HRQOL data were cognitive impairment in 36 (11.9%) patients, dysphasia in 14 (4.6%) patients, communication problems in 5 (1.7%) patients, and hospitalization in 1 (0.3%) patient. When compared with those with HRQOL data, subjects with missing HRQOL data were older (median age, 79.5y vs 71y), had a higher median NIHSS score (14 vs 5), and a lower median Barthel Index score (5.5 vs 20). All these differences were significant (Mann-Whitney U test, P<.001).

Table 2. Changes in Functional Status and HRQOL in the First Year of Stroke

	Measure	3 Months (t3)		6 Months (t6)		12 Months (t12)
		N	Score	N	Change (t6−t3)	N	Change (t12−t3)
	Barthel Index	303	19(15–20)	297	0(0to0)	268	0(0to1)
	IADL	303	1.1(0.1–1.4)	297	0(−0.1to0.0)	268	0(−0.1to0.1)
	MMSE	285	24(19–28)	271	0(−1to2)⁎	251	0(−2to2)
	GDS	263	5(2–10)	254	0(−1to2)⁎	227	1(−1to2)⁎
	WHOQOL-BREF (HK) domain
	Physical	247	56(44–69)	237	0(−7to6)	213	0(−9.5to7.0)
	Psychological	247	56(44–70)	237	0(−7to6)⁎	213	0(−12to6)⁎
	Social Interaction	247	69(56–75)	237	0(−6to0)⁎	213	0(−12to0)⁎
	Environment	247	69(63–75)	237	0(−12to0)⁎	213	0(−12to6)⁎

NOTE. Values are median (interquartile range).

⁎ P<.025 (Wilcoxon signed-rank test with Bonferroni adjustment).

We performed the great majority (87%) of baseline measurements at 3 months in the outpatient clinic at PWH; however, because many stroke patients were discharged from or infrequently followed up in specialist clinics at PWH, 56% and 85% of 6- and 12-month follow-up were performed by home visits. Six of the 303 subjects could not be interviewed at 6 months (3 moved to Mainland China; 2 died; 1 could not be contacted), and 35 subjects dropped out at 12 months (16 died; 5 moved to Mainland China; 5 could not be contacted; 4 refused; 2 were seriously ill; 3 other reasons). When compared with those with complete data, dropouts at 12 months were older (median age, 77y vs 72.5y), had a higher median NIHSS score (8 vs 6), and a lower median Barthel Index score (15 vs 19). All differences were highly significant (Mann-Whitney U test, P<.001). Six patients had another stroke and 10 patients were admitted to nursing homes between the 3-month and 12-month follow-up.

The changes in functional status, mood, and HRQOL of subjects from 3 months to 12 months after stroke are shown in table 2. There was no significant change in Barthel Index and IADL scores, but there was a significant increase in GDS and MMSE scores, and a significant decrease in psychological, social interaction, and environment HRQOL scores at both 6 and 12 months (P<.025 for each).

Multilevel multivariate regression analysis revealed that there was a significant reduction in social interaction HRQOL score at 12 months and in environment HRQOL score at 6 and 12 months, when compared with scores at 3 months (table 3). Reported HRQOL for physical health was significantly lower among women than among men, and HRQOL for a patient’s environment was lower if a patient resided in a nursing home. Other significant covariates of a reduced HRQOL were receipt of comprehensive social security assistance or disability allowance (physical and psychological domains); intake of soft food or tube feeding, and moderate-to-severe pain (physical domain); and exercise of less than 20 minutes a day (psychological and social interaction domains). An increasing GDS score was associated with a reduction in all 4 HRQOL domains, whereas an increasing Barthel Index score was associated with increased HRQOL for physical and psychological domains.

Table 3. Results of Multilevel Multivariate Analysis: Modeling Change in HRQOL Domain Scores

	Independent Variables	Physical HRQOL	Psychological HRQOL	Social Interaction HRQOL	Environment HRQOL
	Follow-up
	6 month vs 3 month				−2.8(−1.2to−4.4)
	12 month vs 3 month			−2.5(−0.7to−4.3)	−1.8(−0.2to−3.4)
	Sex
	Female vs male	−5.0(−2.6to−7.4)
	Residence
	Old-age home vs own home				−5.3(−1.5to−9.1)
	Welfare assistance
	CSSA vs none	−4.0(−0.6to−7.4)	−4.2(−0.8to−7.6)
	Disability allowance vs none	−4.6(−1.4to−7.8)	−3.4(−0.4to−6.8)
	Diet
	Soft diet/tube feeding vs normal	−4.2(−0.8to−7.6)
	Pain
	Moderate/severe vs none	−4.9(−0.9to−8.9)
	Physical exercise per day
	<20min vs ≥1h		−3.1(−0.1to−6.1)	−3.9(−0.9to−6.9)
	Barthel Index	1.2(0.8to1.6)	0.6(0.4to0.8)
	GDS	−1.8(−1.4to−2.2)	−2.6(−2.4to−2.8)	−1.2(−0.8to−1.6)	−2.0(−1.6to−2.4)

NOTE. Values are coefficient (95% confidence interval). Only significant factors are shown; nonsignificant factors were as follows: age, sex, marital status, education, dysphagia, receipt of old-age assistance allowance, mild pain, 20 to 59 minutes of exercise per day, IADL score, MMSE score, and NIHSS score. Abbreviation: CSSA, comprehensive social security assistance.

Among the 213 patients who completed the 12-month follow-up and supplied HRQOL information, 53 (25%), 65 (31%), 58 (27%), and 57 (27%) showed a reduction of HRQOL by at least 8 percentage points for the physical, psychological, social interaction, and environment domains, respectively. In multivariate logistic regression analyses of baseline variables, being married was associated with a decline in physical HRQOL (odds ratio [OR]=2.38; 95% confidence interval [CI], 1.22–4.63; P=.011). A decreasing GDS score was associated with a decline in both psychological HRQOL (OR=.91; 95% CI, .85–.97; P=.004) and environment HRQOL (OR=.91; 95% CI, .85–.97; P=.005). No covariate was associated with a decline in social interaction HRQOL.

Discussion 

This prospective cohort study of patients during the first year of stroke showed that self-perceived HRQOL deteriorated significantly at 12 months in terms of social interactions and living environment. Depression had a negative association with all 4 domains of HRQOL, whereas decreasing disability in basic ADLs (reflected by an increasing Barthel Index score) had a positive association with physical and psychological HRQOL. Female sex, residence in a nursing home, some forms of welfare assistance, restriction to soft food or tube feeding, pain in affected limbs, and little exercise were negatively associated with various aspects of HRQOL.

The strengths of this study were its low dropout rate and the wide range of disabilities among the cohort. We minimized the dropout rate by interviewing patients in their homes if necessary. However, the few patients who defaulted from follow-up were older and more severely disabled than those who completed the study. To minimize the bias toward more favorable outcomes that this disparity would have introduced, we used the statistical method of multilevel modeling. A limitation of the study was that the WHOQOL-BREF (HK) questionnaire and GDS required a degree of intellect and cooperation. Not unexpectedly, patients who could not be assessed for their HRQOL were older and had more severe strokes than those who could. Hence, the number of very disabled patients with stroke who were available for HRQOL analysis was small (32 with a Barthel Index score of <15).

We hypothesized that HRQOL would improve with time, as psychologic, physical, and social readjustment take place. But this was not the case in our cohort. Multilevel analysis revealed that social interaction and environment HRQOL scores were significantly lower at 12 months than at 3 months. This result could not be explained by changes in functional status, because median Barthel Index and IADL scores were unchanged during follow-up. On the other hand, there was a significant increase in depressed mood as reflected by the significant increase in median GDS score from 3 to 12 months (5 vs 6, P<.025). Moreover, the GDS but not the Barthel Index or IADL was associated with these 2 domains of HRQOL. Hence, the decline in HRQOL was probably attributed to psychosocial factors.

The environment HRQOL includes use of transport, access to information, access to health services, leisure activities, financial sufficiency, and home safety. Multivariate regression analysis indicated that patients who showed a significant decline (of ≥8 percentage points) in environment HRQOL tended to be those who had lower GDS scores at 3 months. This finding suggests that initially psychologically well-adjusted patients may increasingly feel constrained when they attempt to reorganize their lives after the stroke. Self-help groups led by trained stroke survivors and social workers may be helpful by promoting mutual sharing of life experiences after stroke and by providing peer support. Nursing home residence was also associated with poorer environment HRQOL. The financial burden from the costs of nursing homes might be an explanation for this association. On the other hand, there is much that the home operators can do to improve on the safety of the living environment, leisure activities offered, and access to information—all without much additional cost.

The decline in social interaction HRQOL suggests that the patients with stroke experienced a worsened personal relationship with family members or friends. Sexual dysfunction, which is very common after strokes,19 might contribute to a strained relationship with a partner; this problem could respond to counseling. It is noteworthy that in the multivariate regression analysis of factors associated with a clinically significant reduction in HRQOL domain scores at 12 months, there was no predictor for a decline in social interaction HRQOL. It is possible that we missed some important variables relevant to this domain of HRQOL, such as stigmatization,20 self-esteem,21 and extent of social networks.6

It is interesting that poor social interaction HRQOL was associated with a lack of physical exercise. The psychosocial benefits of physical exercise among disabled people are well recognized.22 A recent randomized controlled trial showed that home exercise supervised by therapists reduced incidence of depressive symptoms in stroke patients.23 Family caregivers can be trained to supervise physical exercise,24 but compliance is a problem in the longer term. Group exercises may be more effective and enjoyable: a local pilot study25 showed that group exercises supervised by a physiotherapist in a social center setting were effective in improving mobility and stability in elderly people who had been chronically disabled by stroke.

In general, both depression and basic functional status were important determinants of HRQOL, which is consistent with findings of other researchers.5 In this study, the negative effects of depression on HRQOL were more generalized than those of physical disability. Poststroke depression is very common and is associated with impaired functional recovery26 and cognitive impairment.27 Antidepressants for poststroke depression are effective in improving mood, speeding functional recovery, and reducing mortality in the long term.28 Clinicians should therefore actively look for poststroke depression and immediately initiate antidepressant therapy, because there is some evidence that antidepressants can prevent poststroke depression.29 Given the significantly elevated mean GDS score at 1 year, clinicians should remain alert to the possibility of poststroke depression months and possibly years after the stroke event.

Of the 4 domains of HRQOL, the physical domain was the most impaired at 3 months (at 56%). The most modifiable associated factor was pain in the affected limbs. For example, a painful shoulder of a hemiplegic arm is common and may persist for a long time,30 yet proper positioning and handling of the hemiplegic arm is important in the prevention of pain, and electrical stimulation and botulinum toxin injection may offer relief.31, 32 Thalamic pain after a stroke is not uncommon and is difficult to treat. Treatment with drugs such as amitriptyline, carbamazepine, and gabapentin, and acupuncture may be helpful.33 Pain from spasticity may be relieved by botulinum toxin injection.34 Another potentially modifiable factor that was associated with physical HRQOL was the feeding method. The discomfort of tube feeding is well known, and soft-food and puree diets limit the choice of foods. The need for tube feeding and soft-food diets should be regularly reviewed. A local study found that one quarter of nursing home residents who had been tube fed for more than 6 months were found to be able to eat after assessment by a speech therapist and videofluoroscopy.35 In addition, more thought could be given to the presentation of pureed foods.36

Nonmodifiable factors that were linked to HRQOL in the various multivariate analyses included the sex and the marital status of patients. These factors may shed some light about the causes of poor HRQOL after stroke. First, female sex was associated with poorer physical HRQOL. This result was also found in a cohort study of stroke survivors in Spain.5 A possible explanation for our finding is that after retirement, Chinese women have a higher expectation on their physical activity levels because of the burden of domestic duties. More investigation into the observed sex difference in HRQOL is warranted. Second, stroke survivors with spouses had an increased likelihood of declining in physical HRQOL. Perhaps these patients have higher expectation of recovery from physical disabilities because of their concern of the spouse’s burden. Another possibility is that the spouse caregivers overprotect the stroke survivor, thereby making the survivor feel increasingly dependent on care. How spouses view and manage a stroke survivor’s physical disabilities may therefore have a major influence on the HRQOL.

Stroke is associated with financial hardship either directly by the loss of work-related income and costs of formal care (eg, domestic maid, nursing home), or indirectly by loss of income of family caregivers. Consequently, a significant proportion of stroke patients had to receive comprehensive social security assistance or disability allowance for the first time. It is not surprising that financial hardship as indicated by welfare assistance was associated with poorer psychologic HRQOL. However, its association with poorer physical HRQOL suggests that financial constraints might have adversely affected their perceived health, probably by not getting adequate personal care or by limiting access to health care. Relative to Hong Kong, many Asian countries have even less social welfare assistance and access to affordable health care for the poor and disabled. The reduction in physical and psychologic HRQOL in stroke patients in those countries can be expected to be even more significant.

Conclusions 

Environment and social interaction HRQOL may diminish after 1 year of stroke among elderly Chinese patients. Some of the factors that are associated with HRQOL—depression, diet, exercise, and limb pain—are potentially modifiable, whereas others—sex, marital status, and poverty—are not. We recommend that all stroke patients should be followed up by a health care professional for at least 1 year. Apart from the need to control risk factors of recurrent strokes, attending doctors or nurses should focus on treating depression and pain in the affected limbs, and on reviewing the need for a soft-food diet or tube feeding. Group exercises and self-help group activities may help in promoting socialization and reintegration into community life.