The Burden of Stroke Scale (BOSS) provides valid and reliable score estimates of functioning and well-being in stroke survivors with and without communication disorders

https://doi.org/10.1016/j.jclinepi.2003.11.016Get rights and content

Abstract

Background and objective

This study describes the conceptual foundation and psychometric properties of the Burden of Stroke Scale (BOSS), a patient-reported health status assessment designed to quantify the physical, cognitive, and psychological burden of stroke.

Methods

Qualitative research methods were used to develop a 112-item pilot version of the instrument. The pilot version was administered to healthy controls (n = 251) and stroke survivors with (n = 135) and without (n = 146) communication disorders on a single occasion for the purposes of reducing the global item pool, describing the resulting scale properties, examining the dimensionality of the burden of stroke construct, and examining the known-groups validity of the instrument.

Results

Sixty-four items were retained, comprising 12 internally consistent and unidimensional scales. Principal components analysis revealed three second-order factors (Physical Activity Limitations, Cognitive Activity Limitations, and Psychological Distress) comprising the Burden of Stroke construct. Comparisons between groups revealed that stroke survivors reported greater activity limitations and psychological distress on all scales relative to controls, and that stroke survivors with communication disorders reported greater activity limitations on swallowing, communication, cognition, and social relations scales relative to non–communicatively disordered stroke survivors.

Conclusion

These findings support the internal consistency of the BOSS scales, the hypothesized dimensionality of the burden of stroke construct, and the known-groups validity of the instrument.

Introduction

Stroke is one of the leading causes of long-term disability in adults [1], [2], [3]. It is estimated that 4.5 million Americans are currently living with the effects of stroke, and that every year another 590,000 will survive an initial or recurrent event [4]. Traditionally, the measurement of nonfatal stroke outcomes has focused almost exclusively on neurological symptom status and aspects of physical functioning [5], [6]. In recent years, however, considerable progress has been made in quantifying the consequences of stroke on health-related quality of life (HRQOL) [7], [8]. HRQOL refers to the impact of a health condition or healthcare intervention on individuals' perceptions of their functioning in physical, cognitive, and social domains of life and of their psychological well-being [9], [10], [11], [12]. Despite the emergence of instruments designed to quantify the consequences of stroke on these broader dimensions of health, conceptual and psychometric limitations remain in the measurement HRQOL in stroke survivors [13].

One such limitation is the way the psychological and emotional consequences of stroke have been conceptualized and assessed with existing instruments. Psychological or subjective well-being (SWB) is a fundamental component of HRQOL that is distinct from the subjective appraisal of functioning. It is also a construct that has been shown to predict many important health outcomes independent of functional status, including survival length in patients being treated for breast cancer [14], mortality in the elderly [15], and stroke incidence in hypertensive men [16]. Further, more than three decades of scientific inquiry have revealed SWB to comprise three psychometrically distinct dimensions: positive affect, negative affect, and domain satisfactions [17], [18], [19], [20], [21], [22].

Whereas existing measures include scales assessing the frequency with which respondents experience negative mood states, the consequences of stroke on domain satisfactions are rarely examined. Such indicators are not trivial in understanding and describing HRQOL, as patient self-appraisals of functioning have been shown to be paradoxically related to judgments of domain satisfactions [23], [24], and both indicators may influence the focus and outcome of interventions.

A second limitation of existing instruments is the lack of explicit, a priori models of the constructs they purport to measure, the dimensionality of those constructs, and empirical evidence supporting the hypothesized conceptual models. That is, although it is well accepted that HRQOL is a multidimensional latent construct [12], no single universally accepted definition has emerged, and it is frequently operationalized in disparate ways across instruments [25].

Finally, despite the prevalence of stroke-related communication disorders [26], [27] and the impact of communication impairment on quality of life [28], [29], [30], [31], communicatively disordered stroke survivors are rarely represented or described in studies validating HRQOL instruments [7], [8], [32], [33]. This is problematic, because communicatively disordered stroke survivors represent a highly prevalent and known clinical subgroup whose health-related states and outcomes may be substantively different from non–communicatively disordered stroke survivors. Some investigators have tried to address this issue by proxy substitutions for patient responses, but the available evidence suggests that proxies routinely overestimate the severity of patient health states and that the poorest agreement between patient and proxy reports is found for domains assessing cognitive, communicative, and emotional status [34], [35], [36].

This study addresses the afore-mentioned limitations by describing the conceptual framework, scale properties, and factor structure of the Burden of Stroke Scale (BOSS) in a sample of community-dwelling stroke survivors, and by examining the known groups validity of the instrument in healthy controls and stroke survivors with and without communication impairment.

Section snippets

Conceptual foundation

The BOSS is a patient-reported, health-status assessment designed to quantify the consequences of stroke on functioning and psychological well-being. BOSS items, response options, and administration procedures were constructed to address the assessment challenges posed by communicatively impaired stroke survivors. It is conceptually grounded in the World Health Organization (WHO) model of functioning, disability, and health (the ICF) [37] and in Diener's conceptualization of SWB [22]. The

Item development

Literature reviews, expert panels (n = 11), and focus groups with stroke survivors (n = 63), family caregivers (n = 56), and healthcare providers (n = 24) were conducted to generate items for the pilot version of the BOSS. Content analyses of the focus groups resulted in the development of 90 items: 38 items assessing four domains of physical functioning (Mobility, Self-care, Sensory Functions, and Energy and Sleep), 37 items assessing three domains of cognitive functioning (Communication, Cognition,

Sampling

Stroke survivors and healthy controls were recruited from collaborating centers located in Oakland, CA; Phoenix, AZ; Pittsburgh, PA; Tampa, FL; and Salt Lake City, UT. National stroke prevalence data [43] were used to establish sampling targets for gender, race, and presence of communication disorder for a sample size of 281 stroke survivors. The targets were then distributed equally across collaborating centers. These data are available from authors.

Study eligibility criteria for stroke

Sample characteristics

The control sample comprised 136 males and 115 females; 76 were non-Caucasian. The stroke sample comprised 161 males and 120 females; 62 were non-Caucasian. Type of stroke was thromboembolic in 59%, hemorrhagic in 15%, and undetermined in 25% of the sample. Site of lesion was left cortical/subcortical in 34%, right cortical/subcortical in 25%, bilateral in 4%, and undetermined in 37% of the sample. The mean Modified Rankin Scale score was 1.96 (SD = 1.28), and months post onset of stroke varied

Stroke survivors vs. healthy controls

Table 3 gives mean scores, standard deviations, and standard effect sizes for BOSS domain, composite and summary scales for stroke survivors (n = 281) and healthy controls (n = 251). All means were adjusted for reported number of comorbid conditions and annual income level. ANCOVA procedures revealed all group comparisons to be statistically significant [F(1,528) ⩾ 20.78, P < .001], with stroke survivors reporting greater activity limitations, psychological distress, or both on every scale of the

Discussion

This study described the conceptual foundation, scale properties, factorial validity, and known-groups validity of the BOSS in a sample of stroke survivors with and without communication impairment. The BOSS was designed to measure patient-reported activity limitations in physical and cognitive functioning, the psychological distress and participation restrictions associated with such activity limitations, and components of emotional well-being. The conceptual foundation of the BOSS is grounded

Acknowledgments

The authors wish to acknowledge T. Bedirhan Ustun, Somnath Chatterji, Joanne Epping-Jordan, Shekhar Saxena, and Can Celik of the World Health Organization Assessment Classification and Epidemiology Program for their many insightful discussions regarding the conceptual development of the BOSS. The authors also acknowledge the assistance of Lawrence Wexler and Carol Barch, of the UPMC Stroke Institute; Margaret Reidy and Lynn O'Donnel, of the UPMC Rehabilitation Institute; Amy Goda and Mary Ann

References (55)

  • D. Goode

    The national quality of life for persons with disabilities project: a quality of life agenda for the U.S.

  • G.H. Guyatt et al.

    Measuring health-related quality of life

    Ann Intern Med

    (1993)
  • M.W.M. Post et al.

    Quality of life and the ICIDH: towards an integrated conceptual model for rehabilitation outcomes research

    Clin Rehabil

    (1999)
  • H. Schipper et al.

    Quality of life studies: definitions and conceptual issues

  • B.A. Golomb et al.

    A review of health-related quality of life measures in stroke

    Pharmacoeconomics

    (2001)
  • A. Coates et al.

    Australian New Zealand Breast Cancer Trials Group. Prognostic value of quality-of-life scores during chemotherapy for advanced breast cancer

    J Clin Oncol

    (1992)
  • J.M. Mossey et al.

    Self-rated health: a predictor of mortality among the elderly

    Am J Public Health

    (1982)
  • S.M. Agewall et al.

    Stroke was predicted by dimensions of quality of life in treated hypertensive men

    Stroke

    (1998)
  • N.M. Bradburn et al.

    Reports of happiness: a pilot study of behavior related to mental health

    (1965)
  • M.J. Stones et al.

    Structural relationships among happiness scales: a second order factorial study

    Soc Indic Res

    (1985)
  • E. Diener et al.

    The independence of positive and negative affect

    J Pers Soc Psychol

    (1984)
  • R.E. Lucas et al.

    Discriminant validity of well-being measures

    J Pers Soc Psychol

    (1996)
  • P. Warr et al.

    On the independence of positive and negative affect

    J Pers Soc Psychol

    (1983)
  • E. Diener et al.

    Subjective well-being: three decades of progress

    Psychol Bull

    (1999)
  • G.L. Albrecht et al.

    Rehabilitation success: the interrelationships of multiple criteria

    J Health Soc Behav

    (1977)
  • T.M. Gill et al.

    A critical appraisal of the quality of life measurements

    JAMA

    (1994)
  • S.E. Moss

    Neurogenic communication disorders resulting from stroke

    ASHA

    (1993)
  • Cited by (77)

    • Clinical Scales to Assess Patients With Stroke

      2021, Stroke: Pathophysiology, Diagnosis, and Management
    • Determinants of Living Well With Aphasia in the First Year Poststroke: A Prospective Cohort Study

      2017, Archives of Physical Medicine and Rehabilitation
      Citation Excerpt :

      In our data analysis, we were unable to definitively confirm the unidimensionality of the total score, and because the internal validity of each domain was good (Cronbach α; aphasia=.86; participation=.81; environment=.72; personal=.83; life with aphasia is a 1-item domain), we chose to model the factors that affected each of the 5 domains. The explanatory (independent) variables chosen from the literature review and our own qualitative studies included demographics (age, sex, socioeconomic status, and level of education); level of physical functioning after stroke as an indicator of the severity of stroke using the Burden of Stroke Scale9; Western Aphasia Battery Aphasia Quotient8 as a measure of the severity of aphasia; psychological distress using the Hospital Anxiety and Depression Scale10; quantity of aphasia therapy social network convoy model11 as an indicator of social support; and the Successfully Living with Aphasia Rating Scale4 as a 1-item self-report of the temporal aspect of living successfully with aphasia. We included the 1-item rating of successfully living with aphasia as an independent variable to determine if a simple rating scale captured the same information as the 38-item ALA.

    • Clinical Scales to Assess Patients with Stroke

      2015, Stroke: Pathophysiology, Diagnosis, and Management
    • What are the usual assessment practices in adult non-progressive dysarthria rehabilitation? A survey of Australian dysarthria practice patterns

      2019, Journal of Communication Disorders
      Citation Excerpt :

      However, recent evidence suggests that informal ratings may not provide the most accurate information on speech intelligibility due to poor inter- and intra-rater reliability (Hustad, Oakes, & Allison, 2015; Lousada, Jesus, Hall, & Joffe, 2014; Miller, 2013) and the influence of additional factors, such as acceptability and naturalness, on listener judgements (Walshe, Miller, Leahy, & Murray, 2008). Consistent with Miller and Bloch (2017), it is unclear why respondents in the present survey eschewed formal intelligibility assessments with reported diagnostic validity and reliability measures (e.g., Baylor et al., 2013; Bennett, Thomas, Austen, Morris, & Lincoln, 2006; Bloch & Wilkinson, 2009; Doyle et al., 2004; Hartelius et al., 2008; Miller, Noble, Jones, Allcock, & Burn, 2008; Walshe et al., 2009) in favour of informal intelligibility judgements. Given that test availability and time were reported to be key factors influencing assessment choice, it is possible that these formal intelligibility assessments were either unavailable within the clinical setting or were perceived by SLPs to be too time-consuming.

    View all citing articles on Scopus

    This work was supported by the U.S. Department of Veterans Affairs Rehabilitation Research and Development Service, Project Number C2386R, and the Geriatric Research Education and Clinical Center, VA Pittsburgh Healthcare System.

    View full text