Article Text

Download PDFPDF

Validation of the Stroke Specific Quality of Life scale in patients with aneurysmal subarachnoid haemorrhage
  1. Hileen Boosman1,
  2. Patricia E C A Passier1,2,
  3. Johanna M A Visser-Meily1,2,
  4. Gabriel J E Rinkel3,
  5. Marcel W M Post1,2
  1. 1Rehabilitation Centre De Hoogstraat, Utrecht, The Netherlands
  2. 2Rudolf Magnus Institute for Neuroscience, Department of Rehabilitation and Sports Medicine, University Medical Centre, Utrecht, The Netherlands
  3. 3Rudolf Magnus Institute for Neuroscience, Department of Neurology, University Medical Centre, Utrecht, The Netherlands
  1. Correspondence to Dr MWM Post, Rehabilitation Centre De Hoogstraat, Rembrandtkade 10, 3583 TM Utrecht, The Netherlands; m.post{at}dehoogstraat.nl

Abstract

Background and purpose Disease specific quality of life measures have been validated for patients with ischaemic stroke and intracerebral haemorrhage, but not for patients with aneurysmal subarachnoid haemorrhage (SAH). The aim of this study was to validate the Stroke Specific Quality of Life (SS-QoL) scale for patients with SAH.

Methods Cross sectional survey of 141 aneurysmal SAH patients. Construct and criterion validity were studied and various ways to merge the 12 SS-QoL domains into a limited number of subtotal scores were explored. Statistics included assessing score distributions, Cronbach's α, principal components analysis (PCA) and Spearman correlations between SS-QoL and the Glasgow Outcome Scale (GOS), Cognitive Failures Questionnaire (CFQ), Life Satisfaction-9 (LiSat-9) and Hospital Anxiety and Depression Scale (HADS).

Results PCA revealed two components reflecting physical health and psychosocial health with a mutual correlation of 0.73. A ceiling effect was present for 10 out of 12 domains and for the physical component. Internal consistency was good for all 12 domains (α ≥0.80), two components (α ≥0.95) and the total score (0.97). Physical SS-QoL scores showed weak to moderate correlations (0.24–0.32) with the GOS. All SS-QoL scores showed moderate to strong correlations (0.35–0.72) with the CFQ, LiSat-9 and HADS.

Conclusions The SS-QoL is a valid measure to assess quality of life in patients after aneurysmal SAH. Using physical and psychosocial SS-QoL summary scores simplifies the use of this measure without concealing differences in outcomes on different quality of life domains.

View Full Text

Statistics from Altmetric.com

Introduction

Subarachnoid haemorrhage (SAH) from a ruptured intracranial aneurysm accounts for approximately 5% of all strokes,1 occurs at a relatively young age2 and carries a poor prognosis of survival, despite improvements in medical care for these patients.3 The 50–70% of patients who survive an aneurysmal SAH often experience a decrease in their health related quality of life (HRQoL).4 5 Thus far, HRQoL after SAH has only been assessed with generic HRQoL measures. Generic measures can be administered to a broad range of populations. Hence they do not concentrate on disease specific problems. For outcome research, a combination of generic and disease specific measures is therefore recommended.6

The Stroke Specific Quality of Life (SS-QoL) scale is a well known stroke specific HRQoL measure.7 8 The SS-QoL consists of 49 items and results in 12 domain scores and a total score. The SS-QoL has been extensively validated in patients with ischaemic stroke and intracerebral haemorrhage9–11 but not in patients who survived SAH. The effects on HRQoL may differ between SAH and ischaemic stroke or intracerebral haemorrhage as SAH results in diffuse or multifocal brain damage whereas the other types of stroke result more often in focal damage. Whether the SS-QoL is also a valid instrument to investigate HRQoL following SAH is therefore uncertain. Furthermore, its structure of 12 domains is less practical for research into correlates of HRQoL, leaving a choice to use all 12 domain scores as dependent variables or to use the total SS-QoL score with a risk of concealing differences between HRQoL domains. In an earlier study,12 we therefore merged the 12 domains into four dimensions (physical, cognition, social and emotion) to reflect the four basic HRQoL dimensions.13 Hence the objectives of the current study were: (1) to study the construct validity and criterion validity of the SS-QoL in an SAH population; and (2) to explore different options to merge the 12 SS-QoL domains into a limited number of subtotal scores in this population.

Materials and methods

Subjects

This cross sectional study was part of a larger study on HRQoL after SAH.12 We studied all patients who had been treated by clipping or coiling after aneurysmal SAH between January 2003 and July 2005 at the University Medical Centre Utrecht (UMC Utrecht). Patients living in a nursing home and patients with severe comorbidity, reduced life expectancy or an insufficient command of the Dutch language were excluded. All eligible patients who were willing to participate were asked to complete a mailed questionnaire. The medical ethics committee of the UMC Utrecht approved the study protocol.

Measures

Data on demographic characteristics, location and treatment of ruptured aneurysm, and complications after SAH were obtained from the SAH database of the Department of Neurology and Neurosurgery, UMC Utrecht.

Stroke Specific Quality of Life scale

The SS-QoL contains 12 domains with 49 items in total. The domains are: self-care, mobility, upper extremity function, language, vision, work, thinking, family roles, social roles, personality, mood and energy. All items are rated on a 5 point Likert Scale. Higher scores indicate better functioning. Items are averaged to obtain domain scores and a total SS-QoL score. The possible range of all summary scores is therefore from 1 (poor HRQoL) to 5 (good HRQoL). The SS-QoL was previously translated into Dutch using a forward–backward procedure.14

Criterion variables

As a measure of physical functioning at discharge of the hospital we used the Glasgow Outcome Scale (GOS).15

As a measure of perceived cognitive functioning, we used the Cognitive Failures Questionnaire (CFQ).16 The CFQ consists of 25 items measuring self-reported failures of memory, attention, motor function and perception. Items are rated on a 5 point scale. Scores range from 0 to 100 and a high CFQ score indicates poor cognitive function.

As a measure of perceived social functioning, we used the Life Satisfaction questionnaire (LiSat-9).17 It consists of one question about satisfaction with life as a whole and eight questions about satisfaction with life domains: self-care ability, leisure time, vocational situation, financial situation, sexual life, partnership relations, family life and contacts with friends. Each question is rated on a 6 point scale. The total LiSat-9 score is the average of all item scores and has a 1–6 range. Higher scores indicate higher levels of satisfaction.

As a measure of emotion, we used the total score of the Hospital Anxiety and Depression Scale (HADS).18 The HADS contains seven items about depression and seven items about anxiety. Each question is rated on a 4 point scale. Higher scores indicate more emotional problems.

Statistical analysis

SS-QoL score distributions were examined. Skewness was considered present if this statistic was below −1.0 or above 1.0. Floor or ceiling effects were considered present if more than 15% of all respondents chose the lowest or highest possible score.19

Construct validity was examined using several methods. For internal consistency, Cronbach's α and mean inter-item correlations were assessed. Internal consistency requires a Cronbach α coefficient of at least 0.70.19

Spearman correlations between the 12 domains were investigated to assess whether the individual domains measure different constructs. Principal components analyses (PCA) were used to explore possible groupings of the 12 domains into a limited number of components. Selection of the number of components was based on a plot of the eigenvalues of all components (the screeplot) and the number of components that had an eigenvalue more than 1. In addition, a theory driven four dimensional structure was examined.13 In these PCAs, the 12 domains were used as variables instead of the 49 items because of the limited sample size and very good internal consistency of all 12 domains (see results section). Oblimin rotation was performed as we expected correlations between components of HRQoL.

Criterion validity was assessed by examining the degree to which SS-QoL scores were related to scores on domain specific instruments. Spearman correlation coefficients between 0.30 and/0.60 were considered as moderate and correlations exceeding 0.60 as strong.20

As complications following SAH (eg, secondary ischaemia) could cause additional and more focal brain damage, we also performed all analyses including only SAH patients without rebleeding, secondary ischaemia or hydrocephalus. These analyses revealed only marginal differences between the results of this group and those of the total patient group. Therefore, we present only the results of the total patient group.

Results

Population characteristics

Between January 2003 and July 2005, a total of 212 SAH patients underwent aneurysm occlusion by means of clipping or coiling. Of this group, 21 died, eight were discharged to a nursing home, five were living abroad and four had severe comorbidity. Hence a questionnaire was sent to 174 patients. The response rate was 81% (n=141). No relevant differences between patients who responded and those who declined to participate were found regarding demographic and SAH characteristics.12 Table 1 presents the population characteristics.

Table 1

Characteristics of patients with subarachnoid haemorrhage (n=141)

Dimensionality of the SS-QoL

Investigation of correlations between SS-QoL domains (table 2) showed moderate to strong significant correlations between nearly all domains. Expected strong correlations between SS-QoL domains belonging to the same hypothesised dimension were observed but strong correlations were also present between domains that were not expected to be strongly associated, such as ‘mood’ and ‘work’.

Table 2

Spearman correlations between Stroke Specific Quality of Life domains in patients with subarachnoid haemorrhage (n=141)

PCA without constraints revealed two components with eigenvalues exceeding 1, together explaining 69.7% of the variance (table 3). The screeplot (not displayed) showed a clear break between components 1 and 2. Based on this PCA solution, two subtotal scores were computed, reflecting ‘physical’ and ‘psychosocial’ HRQoL. The mutual correlation between both subtotal scores was strong (0.73). A subsequent forced 4 factor PCA resulted in substantial loadings on the first three components only (table 3). Mutual correlations between the four dimension scores were strong, ranging from 0.62, between the physical and emotional dimensions, up to 0.87 between the social and emotional dimensions. These results do not support a four dimensional structure of the SS-QoL, and this structure was therefore not further analysed.

Table 3

Results of principal components analyses with oblimin rotation of Stroke Specific Quality of Life domains in patients with subarachnoid haemorrhage (n=141)

Descriptive statistics of the SS-QoL

We assessed skewness and floor and ceiling effects (table 4). Substantial ceiling effects were observed for 10 of the 12 domains, most strongly for the domains ‘self-care’ and ‘vision’ on which more than half of all respondents chose the highest possible score. A ceiling effect was also found for the physical subtotal score but not for the other summary scores.

Table 4

Descriptive statistics of the Stroke Specific Quality of Life in patients with subarachnoid haemorrhage (n=141)

Internal consistency

Internal consistency was good for all 12 domains, Cronbach's α coefficients ranging from 0.80 to 0.92. Reliability of the two subtotal scores and the total score was excellent (table 4).

Criterion validity

Table 5 shows the correlations between SS-QoL domain scores and the domain specific instruments. Of the 12 domains, only ‘self-care’, ‘mobility’ and ‘upper extremity function’ were significantly, but only weakly to moderately, related to the GOS scores. All but one domain scores showed moderate to strong correlations with the CFQ, LiSat-9 and HADS. Expected strong correlations were found. The strongest correlation between SS-QoL and CFQ was observed for ‘thinking’. The strongest correlation between SS-QoL and LiSat-9 was observed for ‘social roles’. The HADS showed the strongest correlations with ‘mood’, ‘energy’ and ‘personality’ but showed also moderate to strong correlations (0.44–0.61) with all other SS-QoL domains

Table 5

Criterion validity of the Stroke Specific Quality of Life scores in patients with subarachnoid haemorrhage (n=141)

Only the physical subtotal score was significantly correlated with the GOS. Both subtotal scores and total score showed weak or non-significant correlations with the GOS and moderate to strong correlations with the CFQ, LiSat-9 and the HADS.

Discussion

In this study we validated the SS-QoL in an SAH population. Our results showed the SS-QoL to be a valid measure in assessing HRQoL after aneurysmal SAH. The 12 SS-QoL domains showed good internal consistency, which was also found in other types of stroke.7 10 The SS-QoL showed ceiling effects for all domains except ‘thinking’ and ‘energy’. The physical subtotal score also showed a ceiling effect but the psychosocial subtotal score and the total SS-QoL score did not. One explanation may be that patients with SAH experience few physical impairments after SAH. However, more than half of our patients showed suboptimal outcome on the GOS and previous SS-QoL studies, including patients with ischaemic and haemorrhagic stroke, found similar ceiling effects.7 8 10 21 This ceiling effect is a problem from the psychometric point of view and may limit the usefulness of the SS-QoL in clinical practice if the SS-QoL lacks sensitivity for minor physical health problems. However, the mobility scale, for example, includes broad questions such as ‘did you have problems with walking?’, so, if the patient experiences no problems at all, it is not very likely that clinically relevant problems will be missed.

As expected, we found moderate to strong correlations between the domain-specific measures CFQ, LiSat-9 and HADS and corresponding SS-QoL domain scores. In previous studies in patients with other types of stroke, using other criterion measures, for the most part somewhat lower correlations were found.7–10 22 The physical domains showed only weak to moderate correlations with GOS in our study. In our patient group, only three levels of the GOS were used which decreases its discriminative ability. Furthermore, the GOS was administered at discharge from hospital, about 3 years before the other measures in this study. Finally, the GOS is not an ideal reference measure for physical disability. Previous studies found higher correlations of 0.41–0.62 between ‘mobility’ and the SF-36 physical functioning scale but found correlations of only 0.18–0.26 between ‘upper extremity function’ and the Barthel Index,9 11 which is similar to what we found using the GOS. Our results show that the SS-QoL has a stronger focus on emotional and social health than on physical health.

Correlations between the 12 SS-QoL domains were moderate to strong and PCA revealed a ‘physical’ and a ‘psychosocial’ component. The theoretically driven clustering in four dimensions was not supported by the PCA, the main problem being very high correlations between scores on the ‘emotional’ and ‘social’ scales and dimensions. It is not clear how the 12 SS-QoL domains were selected.7 Previous studies in patients with other types of stroke also did not support the 12 domain structure of the SS-QoL.9 22 In these studies, PCA was executed on all 49 items and one study found eight components9 and another study did not specify this number.22 We refrained from these analyses because of a limited sample size. The use of 49 item scores as variables unavoidably leads to a greater number of identified components than was found in our study using 12 scale scores as variables. Our results nevertheless show that a clustering of the SS-QoL scores in less than 12 domains is possible. One total score might be useful as outcome measure in clinical trials but might obscure differences between different aspects of HRQoL. The 0.73 correlation between the physical and psychosocial subtotal score components shows that one component explains about half of the variance of the other. Using the two SS-QoL subtotal scores for physical and for psychosocial HRQoL might be a good compromise between the simplicity and a need to provide a profile of different aspects of health. The very high Cronbach's α values of these components further indicates that it could be possible to develop a short version of the SS-QoL to facilitate its use in a clinical setting and as an outcome measure in clinical studies.

Our patient group was similar with respect to location of aneurysm, neurological outcome, return to work and mean HADS scores to a large earlier Dutch follow-up study, which increases the external validity of our results.23 Some potential limitations of this study should be noted. Firstly, we excluded eight patients living in a nursing home and four patients with serious comorbidity. Including these patients might have resulted in slightly less ceiling effects of the physical health scores. Secondly, although our criterion measures are well known and were used in various diagnostic groups, they have not been validated in populations with SAH either. Thirdly, measuring physical health with a measure other than the GOS and at the same time as the SS-QoL might have revealed stronger associations between both measures. Fourthly, our study sample was too small to perform PCA on item level for which a sample size of at least 300 is recommended.24 Finally, we did not examine other psychometric properties, such as test–retest reliability and agreement between patients and significant others. Further studies are needed to establish these properties in SAH populations.

References

View Abstract

Footnotes

  • Competing interests None.

  • Ethics approval The study was approved by the Medical Ethics Committee of the University Medical Centre Utrecht.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.