Introduction

Over recent decades, medical advances in the care of people with spinal cord injury (SCI) have led to improved rates of survival and life expectancy.1 In turn, emphasis has shifted from survival to life-long follow-up and quality of life (QOL).2 Despite this change in focus, the degree to which people with SCI experience QOL that is equivalent to their able-bodied peers, and the predictors of that QOL remain underexplored.3, 4 The aims of this study were to compare QOL for people with SCI and their able-bodied peers, and to investigate the relationship between QOL and the components of disability (impairments, activity limitations and participation restrictions) across the lifespan.

Materials and methods

This paper presents, a cross-sectional analysis of data from a prospective and longitudinal study examining outcomes across the lifespan for people with SCI. Data were collected using a pre-arranged and guided telephone interview format in which written questions were provided to participants in advance. This format was chosen to enable participation by a large geographically dispersed sample of people with SCI, for efficiency, ease of administration and to minimize missing data.5

Participants

A random sample of 270 individuals with traumatic SCI was selected from archival records kept by the Queensland Spinal Cord Injuries Service (QSCIS). Demographic characteristics of the study sample are given in Table 1. Inclusion criteria consisted of: (1) SCI of traumatic origin, (2) aged over 15 years and less than 55 years at time of injury, (3) discharged to a community setting after hospitalization and (4) an address within the state of Queensland. Individuals who sustained injuries in childhood or older adulthood were excluded, thus reducing the likelihood of participants displaying either growth-related problems or pre-morbid age-related problems. This yielded a potential sample of 1613 people, which was then stratified into six groups according to time since injury (<5, 5–10, 10–15, 15–20, 20–25 and>25 years). To ensure equivalent representation across strata over the past 60 years, participants were selected randomly from each group and located from QSCIS records and telephone directories, contacted by telephone or by mail and invited to participate. The selection process was repeated until 45 participants were recruited to each group. To achieve the total sample of 270 individuals, 690 names were drawn, of whom 179 could not be located, 64 were found to be deceased, 76 declined to participate, 28 had relocated interstate and 60 were excluded as their motor and sensory function was reported as normal. Those who declined mostly noted insufficient time due to work or family commitments.

Table 1 Demographic characteristics of the study sample at study commencement (n=270)
Full size table

As reflected in Table 2, excluding those who were deceased or those with no residual neurological deficit, the final study sample was comparable to the non-participant sample and the QSCIS population on age, age at time of injury, gender and place of residence. However, the study sample contained a greater percentage of people with tetraplegia and with complete injuries than the non-participant sample and the QSCIS archival population.

Table 2 QSCIS archival record: comparison between the study sample, those selected but who did not participate and the total QSCIS sample
Full size table

Measurement

The survey questionnaire consisted of a core of reliable and valid measures of QOL and disability, suited to an SCI population and telephone administration.

Quality of life

The World Health Organization Quality of Life Assessment Instrument (WHOQOL)-Bref,6 a self-report questionnaire used to measure perceived QOL over the past 2 weeks was the key measure selected as its psychometric properties are excellent for a number of disability groups and it is suitable for measurement of QOL in people with SCI.7 It profiles QOL across four domains; physical health (seven items), psychological health (six items), social relationships (three items), environment (eight items) and two extra items scoring overall perception of QOL and health. Respondents rate each item on a five-point Likert scale. The summed scores within each of the four domains range from 0 to 100, with a higher score denoting higher QOL. The Analysis was performed at three levels: a basic level using the single item ‘how would you rate your quality of life’; an intermediate level using the WHOQOL-8, calculated by summation of eight items contained in the WHOQOL-BREF and a higher level using the four-domain scores (physical, psychological, social relationships and environment).8

Disability

Measurement of disability was based on the International Classification of Functioning, Disability and Health framework.9 Two impairment measures (primary and secondary), one activity measure and one participation measure were chosen to reflect the three components of disability.

Primary impairments were measured according to neurological level and completeness of injury, based on the American Spinal Injury Association (ASIA) classification system,10 with scores re-ordered so that higher scores reflecting more function. That is, a score of 1 denotes functionally complete tetraplegia (Tetra ASIA ABC), 2 functionally complete paraplegia (Para ASIA ABC) and 3 SCI at any level (all ASIA D) with functional neurological sparing such that ambulation is typically possible.

Secondary impairments were recorded using the Secondary Conditions Surveillance Instrument (SCSI),11 in which participants report their experience and the severity of 40 different problems over the past 12 months on a four-point scale. For each participant, the sum of ratings across all conditions (a measure of overall severity of secondary conditions) was calculated.

Activity limitations were monitored using the motor subscale of the Functional Independence Measure (mFIM), universally recognized for measuring function and previously utilized in SCI and ageing research. It consists of 13 items that assess the degree of independence across self-care, sphincter control, mobility and locomotion. Although not suited to clinical interpretation, it covers functional issues relevant to SCI, and higher scores denote higher levels of function. The validity and reliability of telephone administration of the mFIM with people with SCI has been established.12

Participation restrictions were assessed using the community integration measure (CIM), a 10-item, client-centred measure that assesses words of participants with respect to personal attitudes, perceptions and beliefs regarding their connection with their community. A single summary score (between 10 and 50) is derived. The CIM is a psychometrically sound measure when used with people with SCI.13

Procedure

Each telephone interview was arranged and conducted by two physiotherapists with experience in the field of SCI and training in the administration of the interview. Upon receiving verbal consent, an interview time was agreed (daytime or evening) and interview questions, participant information forms and participant consent forms were mailed. Participants were asked to sign and return forms using a reply-paid envelope prior to the first interview. Ethical approval for this project was granted by the Princess Alexandra Hospital Ethical Review Committee, Brisbane, Australia.

Analysis

Data were managed using Microsoft Access Database Management System and analysed using SPSS V13 software. Data quality was high with the frequency of missing item scores less than 10 for the entire sample. Missing data were dealt with using mean substitution for computed variables and pair-wise deletion for single-item variables.

Differences in QOL between people with SCI and the Australian norm were identified using a two-sample t-test. Analysis was performed using the QOL single-item score and at the higher level using the four-domain scores. Effect sizes were calculated for the single-item QOL and for all domains by dividing the mean difference in scores by the s.d. of the scores for the Australian norms. To interpret effect size, the suggestions of Cohen14 were used (0.2–0.5, small; 0.5–0.8, moderate and >0.8, large). Differences between age groups (20s, 30s, 40s, 50s and 60 and over) and time since injury were determined using analysis of variance. The relationship among QOL, the three components of disability (impairments, activity and participation), age and time since injury were examined using univariate linear regression. To determine the relative importance of the components of disability on QOL while taking into account age and time since injury, a multiple linear regression (simultaneous) was utilized for the single-item score, the WHOQOL-8 and each of the four domains of QOL. Measures of QOL were treated as the dependent variable and measures of disability, age and time since injury were treated as the independent variables. Two dummy variables were used to enter neurological status into regression analyses. Significance level was set at <.01 to minimize the probability of making type-1 errors.

Results

On average, participants in this study were found to have significantly lower QOL than the Australian norm for the single-item QOL and for all domains15, 16 (Table 3). The differences in the physical domain (1.00) reflected a large effect size whereas the single-item QOL (0.75) and the social relationships domain (0.53) showed moderate effect sizes, the psychological domain (0.50) had a small–moderate effect size and there was only a small effect size within the environment domain (0.36). When compared for each age group, the difference was significant for all domains, with the exception of the 60s and over group that differed significantly for the physical domain only (Figure 1; Table 4). When QOL was compared between people with SCI of different ages and time since injury, no significant differences were detected.

Table 3 QOL: comparison between study sample and Australian norm
Full size table
Figure 1
figure 1

Quality of life (QOL): comparison between spinal cord injury (SCI) and population norms across age groups. *Significant at <0.05, **significant at <0.01, ***significant at <0.001.

Full size image
Table 4 QOL across age groups for SCI compared to the Australian norm
Full size table

Relationship between QOL and disability across the lifespan

All measures of QOL were associated with SCSI (β=−0.28 to −0.68) and CIM (β=0.29–0.45) but only the physical (β=0.30) and environment domains (β=0.21) were associated with mFIM. No measures of QOL were associated with age, duration of injury and neurological status (Table 5).

Table 5 Bivariate and multivariate relationships between disability and QOL across the lifespan
Full size table

The two factors that accounted for the variance in QOL, when taking into account age and time since injury, were secondary impairments and community integration. In Table 5, regression coefficients, semi-partial correlations and percentage of unique variance are presented. On the basis of the WHOQOL-8, secondary impairments demonstrated the highest semi-partial correlation (sr) among all of the variables (sr=0.51, P<0.001) and were responsible for more than 32% of the unique variance in QOL (WHOQOL-8) and the single largest independent predictor of QOL, irrespective of the contributions of other factors that influence QOL. The CIM also demonstrated a significant semi-partial correlation (sr=0.31, <0.001) and was responsible for 11% of the variance indicating that it is also a strong predictor for QOL. Figure 2 illustrates the comparative importance of each factor entered into the regression equation expressed using β-weights that represent the change in QOL (WHOQOL-8) associated with a one unit change in the individual component of disability with all other factors held constant.

Figure 2
figure 2

Relationships between quality of life (QOL) (World Health Organization Quality of Life Assessment Instrument, WHOQOL-8) and disability: comparative importance of regression analysis factors expressed using β-weights. **Significant at <0.01.

Full size image

Discussion

Consistent with previous reports, QOL for people with SCI in Queensland was significantly lower than the Australian norm.2, 17 QOL was significantly associated with all components of disability with the exception of primary impairments (level and completeness of injury). The single most important predictor of QOL was the presence of secondary conditions whereas the second most important predictor was the extent of societal participation.

Some methodological limitations of the current study must be noted. First, restricting inclusion to people who could be contacted and who had access to the telephone may have led to some selection bias. In addition, some potential participants may have lacked the time to participate due to work or family responsibilities. For some, age-related changes may have occurred that were expected and so were not viewed negatively. Changes in perceived QOL may have resulted from ‘response shift’ or a re-framing of expectations or adapting to life events hypothesized to impact on QOL.18 It is also likely that those potential participants who had died would have had a very different ageing experience from the study participants.10 Finally, it is important to acknowledge that these results cannot be generalized to the SCI population as the study sample was drawn from incomplete archival records of people from only one geographical area. In addition, the sample consisted of a greater percentage of people with a higher level of neurological impairment when compared to those who chose not to participate and to the overall sample from which they were drawn. Further investigations with a number of samples from different geographical areas in Australia will be required before generalizations to the SCI population can be made.

Consistent with other studies,2, 17, 19 findings indicate that QOL for the average person with SCI is lower than the general population norm, irrespective of age or time since injury. Although the differences in QOL were statistically significant, interpreting the clinical significance of these findings is perhaps more challenging and inherently subjective.20 Effect sizes may be taken to reflect the clinical significance of differences across domains. The largest effect size is for the physical domain, which is not surprising given the physical nature of SCI. Moderate effect sizes evident for the social relationships and psychological domains highlight that these domains of QOL remain important to consider in the rehabilitation context. The smallest effect size for the environment domain suggests that individuals may have made significant environmental modifications to address potential difficulties in this area, although this proposition requires further investigation. The findings highlight the central importance of QOL following SCI, as both a goal and outcome of rehabilitation, and support the need to identify how hospital and community rehabilitation services can best address QOL issues.21

Findings with respect to the relationship between QOL and disability were also consistent with previous studies,4, 22 providing confirmation that the experience of SCI in Australia is comparable to other Western nations. QOL was related to impairments (secondary but not primary), activity limitations and participation restrictions, irrespective of an individual's age or time since injury.

Secondary impairments (for example neuropathic pain, urinary tract infection) were found to have the greatest impact on QOL, in keeping with previous studies.22, 23 Although significant attention has been given to the prevention and management of secondary conditions to prevent hospital re-admission,24 these findings underline the need for continued efforts to reduce both the incidence and the impact of these conditions on QOL.

Societal participation was also a significant predictor for QOL, as has been previously reported.25 Of concern, however, is that many rehabilitation services continue to focus their day-to-day service delivery on impairments and activity limitations with less emphasis on participation.21 These findings prompt the need for both hospital and community rehabilitation services to place greater emphasis on assessment and intervention related to social participation, particularly on enhancing factors that facilitate participation and overcoming barriers that impede participation.

Conclusion

QOL for people with SCI was found to be lower than for their able-bodied peers, confirming the need to more effectively address QOL as a goal for hospital and community rehabilitation services. Rehabilitation services must employ improved methods to address the most common and most disabling secondary conditions while at the same time maintaining the focus on functional attainment and providing greater attention to enabling participation. Research is required to identify and investigate the individual factors that influence each of these dimensions of disability.