The impact of patient self assessment of deformity on HRQL in adults with scoliosis

Background Body image and HRQL are significant issues for patients with scoliosis due to cosmetic deformity, physical and psychological symptoms, and treatment factors. A selective review of scoliosis literature revealed that self report measures of body image and HRQL share unreliable correlations with radiographic measures and clinician recommendations for surgery. However, current body image and HRQL measures do not indicate which aspects of scoliosis deformity are the most distressing for patients. The WRVAS is an instrument designed to evaluate patient self assessment of deformity, and may show some promise in identifying aspects of deformity most troubling to patients. Previous research on adolescents with scoliosis supports the use of the WRVAS as a clinical tool, as the instrument shares strong correlations with radiographic measures and quality of life instruments. There has been limited use of this instrument on adult populations. Methods The WRVAS and the SF-36v2, a HRQL measure, were administered to 71 adults with scoliosis, along with a form to report age and gender. Preliminary validation analyses were performed on the WRVAS (floor and ceiling effects, internal consistency and collinearity, correlations with the SF-36v2, and multiple regression with the WRVAS total score as the predictor, and SF-36v2 scores as outcomes). Results The psychometric properties of the WRVAS were acceptable. Older participants perceived their deformities as more severe than younger participants. More severe deformities were associated with lower scores on the Physical Component Summary Score of the SF-36v2. Total WRVAS score also predicted Physical Component Summary scores. Conclusion The results of the current study indicate that the WRVAS is a reliable tool to use with adult patients, and that patient self assessment of deformity shared a relationship with physical rather than psychological aspects of HRQL. The current and previous studies concur that revision of the WRVAS is necessary to more accurately represent the diversity of scoliosis deformities. Ability to identify disturbing aspects of deformity could potentially be improved by evaluating each WRVAS items against indicators of pain, physical/psychosocial function, and self image from previous measures such as the SRS, SF-36 or BSSQ-deformity.


Background
Outcomes from psychosocial and health related quality of life (HRQL) studies indicate that body image is a complex and significant issue for patients with scoliosis and their clinicians [1,2]. According to the body image literature, medical conditions threaten the stability of patient body image via changes to bodily sensation, functioning and appearance [3]. In particular, disfigurement or deformity can promote a negative self image within the individual, who may also experience difficulties with social interaction due to potential adverse reactions from others as a result of the visibility of their condition [4]. As scoliosis is rarely life threatening, the clinician's decision to perform scoliosis surgery on adolescents hinges on current and prospective spinal deformity, with patient HRQL and surgical considerations performing auxiliary roles in the decision making process [5]. Cosmetic issues and physical symptoms are the key indicators for scoliosis surgery in adult patients, even though further curvature progression is extremely unlikely [1]. However, clinical assessments of scoliosis correlate poorly with patient self perceptions of deformity and self reported HRQL [5,6]. Assessment of body image and factors likely to influence body image (HRQL) is important for scoliosis patients, as difficulties in these areas may have an adverse affect on treatment compliance and satisfaction in adolescence, and limit psychosocial functioning in adult life [7][8][9].
Recent attempts to map radiographic and treatment variables to HRQL and body image outcomes in the literature are summarized in Table 1 and Table 2. Earlier HRQL and psychosocial studies are summarized elsewhere [2]. Few consistencies emerge in the literature, and factors such as age, gender and psychological health confound the relationship between clinical measures of scoliosis deformity and body image or HRQL outcomes [7,10,11]. It can be seen that for adolescent patients treated conservatively, brace wear exerts a greater impact on body image and HRQL than the deformity itself [12]. Correlations between radiographic measures and self reported outcome measures range from mild to moderate amongst adolescent surgical candidates, with a tendency for stronger correlations in the Self Image, Function and Pain domains HRQL, and poorer outcomes amongst patients with thoracic curvatures [13][14][15][16]. With the exception of saggital balance, radiographic measures were shown to be even poorer indicators of body image and HRQL outcomes in adults [10,17,18]. The nature of scoliosis deformity is subject to age graded changes, and age itself is associated with poorer HRQL regardless of disease status. Psychosocial studies of adult scoliosis patients have also revealed limitation in social and intimate relationships due to physical difficulties in participation, fear of injury or self consciousness [9,17,19,20].
Assessment of body image in scoliosis patients has been limited to written questions about perceptions of attractiveness in clothing or bathing suits, satisfaction with the body or back, and psychosocial distress as a result of deformity or brace wear [17,19,21,22]. However, few attempts [5,17] have been made to qualify which aspects of deformity are the most distressing for patients. Such information would be useful for clinical decisions, such as whether or not to perform thoracoplasty in addition to spinal fusion, or whether to recommend conservatively treated patients for surgery. The most concentrated effort in this area was the development of a visual analogue scale to quantify patient self assessment of deformity by Sanders et al [23].
Known as the Walter Reed Visual Assessment Scale, the scale features seven items which address visual aspects of scoliosis including: body curve, head pelvis, rib prominence, shoulder level, flank prominence, scapula rotation and head rib pelvis (cf [23][24][25]). Each item consists of five illustrations scaled to indicate worsening deformity via higher scores. The WRVAS is not a body image scale as such. Rather, it is intended to assess the patient's perception of their deformity without cognitive or emotional connotations [23]. Table 3 summarizes the findings of previous studies using the WRVAS. Subsequent studies following initial development have demonstrated stronger and more consistent correlations between the WRVAS and HRQL outcomes than the radiographic studies outlined in Tables 1 and 2. This suggests that the WRVAS is a more accurate reflection of the impact of scoliosis deformity on patient body image and HRQL than radiographic indicators. As with the HRQL measures described in Tables 1 and 2, an attempt to map WRVAS outcomes to radiographic indicators yielded inconsistent findings. Assessment of patient perception of scoliosis deformity provides information unique to radiographic data, and due to its clinical relevance further investigation is warranted.
One notable omission in studies utilizing the WRVAS is the evaluation of WRVAS scores against the Short Form Health Survey (SF-36), version 2. The current study will involve the administration of the WRVAS and Short Form Health Survey, version 2 [SF-36v2] to a sample of adult scoliosis patients in order to determine the impact of patient self assessment of deformity on a HRQL instrument widely used in and populations, and further validate the WRVAS.

Methods
Cross sectional methodology was used to determine the reliability and construct validity of the WRVAS. Support groups, orthopaedic specialists and a large metropolitan university were approached in 2004 and 2005 in an HRQL/Other symptoms: Non surgically treated group had a higher incidence of surgical risk factors (heart disease, overweight). General Health as measured by SF-12 was poorer in non surgical group. Surgical patients had a higher incidence of back/leg pain, and lower scores on the Role Physical and Bodily Pain domains of the SF-12. S urgical patients were more likely to report that the shape of their back had changed over the last 10 years and that they were very unhappy with the shape of their back. They also rated the appearance of their trunk as fair (compared to good amongst non surgical patients), and were more likely to state that their back limited personal relationships. Deciding factors for not selecting surgery were: older age, higher Body Mass Index Deciding factors leading to surgery were: lower SRS Self Image scores, larger thoracic curvature, greater back pain (ODI) Glassman et al [10] 298 adult patients (84% female, age range     Curvature pattern: Patients with King Type II/III curvatures also had a lower BMI and were lower in menarchal status. They were more likely to report neutrality or dissastifaction with surgical outcomes. Satisfaction with surgery: Neutrality/dissatisfaction with surgery was related to lower scores on the OFFER Self Image Q prior to surgery, and lower scores on the Body Self Relations Q post surgery (especially low satifaction with mid/upper torso).  attempt to obtain equal samples of males and females over 18 years of age with scoliosis. The total sample included 13 males and 63 females. Participants completed a questionnaire package which included a form to report age and gender, the Short Form Health Survey, Version 2 (SF-36v2), and the WRVAS.

SF-36v2 questionnaire
The SF-36v2 is an updated version of the SF-36. Currently, the SF-36 is the most popular health related quality of life (HRQL) instrument used on adult populations with scoliosis [26,27]. The second version has been updated to simplify the layout, wording and response formats to minimize cultural bias [28,29]. Like its predecessor, the SF-36v2 is composed of eight subscales Physical Functioning (PF), Role Emotional (RE), Role Physical (RP), Bodily Pain (BP), Social Functioning (SF), Mental Health (MH), Vitality (VT) and General Health (GH). These subscales can be summarized into Physical Component Summary (PCS) and Mental Component Summary (MCS) scores. All scales are reported as T scores, which correspond to a mean of 50 [29].

WRVAS questionnaire
As described earlier, the WRVAS is a seven item scale designed to evaluate patient perception of spinal deformity. Scores are obtained by totaling responses to each of the seven questions [24]. For each item the minimum possible score is 1 and the maximum is 5. The lowest possible score for the total is 7, while the highest possible total score is 35.

Statistical Analysis
Previous validation studies of the WRVAS set precedence for statistical report [24][25][26]. As such, the means and standard deviations for each WRVAS item and the total WRVAS score, along with floor and ceiling effects were reported.
Internal consistency was evaluated via Cronbach's alpha to assure that all items measured a common underlying construct. According to Nunnally [30] a Cronbach's α of 0.7 is considered acceptable, while a value of 0.8 is 'good', and a value of 0.9 is 'excellent.' Collinearity statistics in the form of tolerance and the variance inflation factor were also examined to identify multicolinearity amongst items in the WRVAS. Tolerance values greater than 0.1 and less than 10 indicate that all items of the WRVAS are unique in their contribution to the measurement of scoliosis deformity, in that none of the items are strongly intercorrelated and redundant [24,31].
Construct validity was addressed by correlating each item of the WRVAS and total WRVAS score with each scale of the SF-36v2. Two multiple regression analyses were also conducted to determine the capacity of the WRVAS to predict Physical and Mental Component Summary scores of the SF-36v2. Specifically, total WRVAS score served as the independent variable in each analysis, whilst Physical and Mental Component Summary acted as dependent variables. Age and gender were also examined in the correlations and multiple regressions for the purposes of statistical control.

Results
Five female participants were excluded from the sample due to insufficient questionnaire completion. This resulted in a final sample of 13 males and 58 females or a ratio of 1:4.5, which is a similar ratio to another published study [27], and reflects the clinical population of scoliosis patients [32]. Missing data was substituted for participant subscale means in 13 cases (5 male), where at least four items of the WRVAS or half of the items of a subscale within the SF-36v2 was completed. The mean (standard deviation) participant age was 33 (12.7) years, and age ranged from 17 to 66 years.
The mean (standard deviation) of each WRVAS item and total WRVAS score is reported in Table 4. The percentage of participants who scored the maximum and minimum scores on each scale are also presented in Table 4. Total WRVAS scores were comparable for males and females. Statistically significant correlations were found between Age and Body Curve (r = .327), Head Pelvis (r = .255), Rib Prominence (r = .309), Flank Prominence (r = .351), Scapula Rotation (r = .297), Head Rib Pelvis (r = .342), and total WRVAS score (r = .339). This indicated that older participants assessed their deformity to be more severe than younger participants.

Reliability (internal consistency)
The Cronbach's alpha statistic was 0.925, which is indicative of excellent internal consistency. Furthermore, there was no evidence of collinearity within the seven items (tolerance 0.213 to 0.429, VIF 2.33 to 4.704).

Construct validity
Correlations between each item of the WRVAS scale and total score and domains of the SF-36v2 are displayed in Table 5. Where significant correlations were found, higher scores on the WRVAS were associated with lower scores on the SF-36v2, which indicates a lower quality of life.

Multivariate analysis
The combination of independent variables: total WRVAS score, age and gender, predicted 25.2% of the variance in Physical Component Summary score (F = 7.514, p < 0.01, R = .502, R 2 = .252). An examination of the independent variables revealed that total WRVAS score and gender con-

Discussion
The aim of the current study was to investigate the psychometric properties of the WRVAS on an adult sample of patients with scoliosis. Floor effects were present for all items of the WRVAS, indicating that respondents may not have been aware of minor scoliosis deformity. The internal consistency was excellent (Cronbach's alpha 0.925) and indicated strong correlations between all items of the WRVAS. However none of the WRVAS items were redundant, as statistical testing revealed no evidence of colinearity. The psychometric properties obtained in the current study including floor and ceiling effects, internal consistency and collinearity were comparable to a previous study by Pineda et al [24] utilizing a predominantly adolescent sample. This suggests that the WRVAS is similarly reliable when administered to adult and adolescent samples.
Analysis of construct validity revealed that patient self assessment of deformity shared a stronger association with physical aspects of quality of life. Items of the WRVAS shared a consistent negative correlation with the Physical Functioning, Vitality and General Health subscales of the SF-36v2, and the Physical Component Summary score. Construct validity outcomes of the current study differ from those of Pineda et al [24], which were that the WRVAS items demonstrated a stronger and more consistent relationship with the Mental Health and Self Image domains of the SRS-22, compared to Pain and Function. One possible explanation is the difference in age group for each study, and normative body image concerns for adolescents compared to older adults. A signifi-  [33]. This assertion is suggested in the adult scoliosis literature, with patients reporting greater limitation in physical aspects of HRQL compared to population norms and control groups more consistently than disruptions in psychological HRQL. Although psychosocial studies demonstrate that appearance is still a valid body image issue for adults with scoliosis, most patients seem to find physical health problems associated with scoliosis more limiting in their daily lives [19].

Limitations
As previous authors have noted [23,25] the current version of the WRVAS appeared to possess limited face validity in instances where the patient's condition differed from the item depictions of a right thoracic curvature. In the current study, eleven participants made notes referring to their own scoliosis stating how their conditions differed from the illustrations. Of these respondents, two stated that their curvatures were to the left, two reported lumbar curvatures, two reported double major curvatures, three reported that their results were pre or post surgical, and three stated that they had other conditions connected to their scoliosis.
There were two further limitations of the current study that have already been acknowledged. Firstly, there was a small sample size in general, and an unequal distribution of males and females. It is likely that gender is a salient issue in the measurement of body image in adults with scoliosis, given the practically significant results obtained despite the small sample size.
Secondly, scoliosis qualifiers such as curve magnitude and treatment type were not collected from the sample. While this data would have provided more information and enabled for greater statistical control of possible confounding variables, the results obtained still provided useful information with this omission as the purpose of the study was to evaluate patient perceptions of deformity and HRQL. Previous studies have demonstrated a strong correlation between curve magnitude and total WRVAS score in younger participants [23,24,34,35]. However it is unlikely that radiographic indicators would have been useful for the adult sample assessed in the current study, as HRQL demonstrates a poor relationship with radiographic measures in adulthood, with the exception of sagittal balance.

Conclusion
The results of the current study confirm that the WRVAS is a psychometrically valid tool for use with adult scoliosis populations. These findings add to the complexity of body image data in the scoliosis literature, as it was suggested that physical health factors such as pain and functional capacity are especially salient to body image amongst adult patients. Outcomes of the current study indicate that there is scope for improvement of the WRVAS by increasing the scope of curvature patterns represented, and incorporating items salient to age factors such as kyphosis and lordosis. Furthermore, more comprehensive assessment of HRQL and body image could be achieved by asking patients to evaluate aspects of deformity measured by the WRVAS against body image and HRQL outcomes. Such outcomes include social functioning, satisfaction with appearance, attractiveness, pain, physical functioning and emotional functioning. Measures of these variables could be adapted from previous questionnaires including the SRS, SF-36 or BSSQ -Deformity instruments, all of which have been previously validated on scoliosis populations.