For measuring of Cobb angle, vertebral rotation and trunk listing, computer-assisted digital radiographic measurement is used in the authors’ institution, the Picture Archiving and Communication System (PACS). Digital imaging has many advantages in terms of convenience and the ability to adjust contrast, brightness and magnification, leading to increased accuracy of measurements in comparison to manual methods . Kuklo et al.  commented that digital measurement improves measurement precision and shows good correlation with manual measurements for the majority of AIS parameters. Shea et al.  reported that the 95% confidence interval for intraobserver variability was 3.3° for manual method but 2.6° for computer-assisted measurement. Srinivasalu et al.  also reported the 95% confidence interval for intraobserver and interobserver variability to be 1.3° and 1.26°, respectively. More importantly, Gstoettner et al.  reported the main source of error to be the definition of end vertebrae. When the variability in selection of the end vertebrae was eliminated, the amount of actual error in the measurements among the examiners was relatively small . The source of error may be reduced for the computer-assisted measurement because the software measures the angle automatically after drawing lines through the endplates of the end vertebrae. Besides, most studies highlighted advantages of using the torsiometer in measuring vertebral rotation. In examining interobserver and intraobserver errors, Weiss  reported the intraobserver error as 1° and interobserver error as 3°. Furthermore, a study by McLean et al.  showed that 95% of the second measurements of trunk listing by the plumb line method would be expected to be within 4 mm of the first measurement. However, no previous studies measuring trunk listing by means of PACS have been documented. In this study, the author as one examiner measured all subjects’ spine radiograph. The 95% confidence intervals for intraobserver variability in Cobb angle, vertebral rotation and trunk listing measurements were 2.8°, 2° and 3.2 mm, respectively, upon Visit 1, 3.2°, 1.8° and 4.2 mm, respectively, upon Visit 2, and 3.0°, 1.6° and 4.5 mm, respectively, upon Visit 3.
Orthosis effectiveness measure
Landauer et al.  suggested that an initial correction of more than 40% and good compliance had significant effects on outcome. To explore its relationship, in-brace correction was categorised as 40% or above and below 40%; and bracing hours was also divided into three groups: from the least compliance to the most compliance, for analysis. No significant difference was detected between the three categories of bracing hours on in-brace correction, but for the least compliant group of patients, the in-brace correction tended to be less than 40%. This implies that the more compliant a patient is, the greater will be the in-brace correction. Besides, Kinel et al.  showed wearing the Cheneau brace a minimum of 16 hours per day, resulted in less clinical deformity than resulted from non-treatment. The study of Rahman et al. , demonstrated that highly compliant patients (85% compliance) showed no curve progression at the end of the treatment, whereas poorly compliant patients (62% compliance) showed curve progression of more than 6°. Although Weiss and Rigo  reminded that it should aim for an in-brace correction of more than 40%, not all curves can be corrected to the same extent. As Landauer et al.  commented, compliant patients with high initial correction can expect a final correction of around 7°, whereas compliant patients with low initial correction may not see a change in the degree of curve, and poor compliance is always associated with curve progression. While compliance is always an issue that needs to be dealt with, the study result suggested that its effect on in-brace correction that should never be underestimated. As suggested by Kim et al. , correction of the curve should be maximised in the brace with careful fitting and adjustment of the pads by an experienced orthotist. Best practice should aim at the best in-brace correction and at the same time the best possible comfort for the patient to foster compliance. The better the in-brace correction, the better the end result. In fact, in-brace correction is negatively related to curve magnitude . Poor results can be due to poor bracing and this could be verified through in-brace radiographs to assess the obtained correction. Poor results can also be due to improper management of the patient, a factor that can ultimately influence compliance . With close monitoring of the patient’s compliance and in-brace correction, a successful treatment outcome can be attained.
A significant difference in in-brace correction was observed between patients with different curve patterns. It was predicted that those with thoracic curve patterns would reach in-brace correction of less than 40%. The correction effect, as reflected by the improvement in Cobb angle, was also shown to be better for the thoracolumbar curve patterns in comparison to the thoracic curve patterns, as detected in Visits 2 and 3. These findings were consistent with those reported by Weiss et al.  that the correction effect was highest for the lumbar and thoracolumbar curve patterns. It was explained that thoracic curve is difficult to correct in comparison to thoracolumbar curve as a result of the anatomical structure where it articulates with the ribs to form the rigid rib cage.
Previous studies have shown that the evaluation of orthosis effectiveness has mainly focused on the correction of Cobb angle after treatment. In this study, other relevant parameters such as vertebral rotation and trunk listing were measured. As mentioned by Perdriolle et al. , other than Cobb angle, measurement of vertebral rotation is also significant in the prognosis and treatment of scoliosis curves. The study revealed a significant positive correlation between in-brace Cobb angle and in-brace vertebral rotation in Visit 2 as well as Visit 3. This finding was consistent with the comment made by Leathermann and Dickson  that vertebral rotation increases with increases in Cobb angle and is a reflection of the severity of the deformity.
Regarding bracing compliance, the subjects were divided into three groups for analysis: group 1 (0–8 hours), group 2 (9–16 hours) and group 3 (17–23 hours) in this study. This was consistent with the design of the log sheet. The subjects logged their brace wearing time as 0000–0800 hours, 0800–1600 hours, or 1600–2400 hours. For a comprehensive review, subjects were asked to note the hours they did not wear the brace, such as while bathing and exercising. From the log sheets, it was not difficult to observe the brace wearing patterns of the subjects and therefore their bracing compliance. For example, those belonged to group 1 most often wore their brace at nighttime. These subjects explained that they tried to avoid the brace during the daytime because they did not want their classmates or friends know that they needed to wear a brace. Hence, they could wear the brace only when they were at home or sleeping. This showed a deep psychological effect as reflected by the poorer QoL score.
It was predicted that the difference in TAPS mean score between Visit 1 (pre-brace) and Visit 2 of the least compliant group (0–8 hours) would be significantly (−0.70) less than that of the most compliant group (17–23 hours). The effect of brace wear was obvious. It may be that patients in the least compliant group were aware they were not compliant and therefore perceived their trunk as rather deformed by the time of Visit 2. To the contrary, the patients in the most compliant group might have reckoned that they were compliant and so perceived the appearance of their trunks as improved by the time of Visit 2. This implies that the patients more or less believed in the effect of brace wear. As highlighted by Borders , one of the important factors that predicts compliance is the belief patient has in the treatment outcome. However, it must be noted that although the least compliant group might have believed in the effect of brace wear, they did not comply with it. Some psychosocial issues may be hidden behind this behaviour and they need to be revealed. As noted by Hawes , psychological issues have previously been shown to influence compliance in brace wear. The perception of body image and that of the trunk deformity are complementary . Self image is decreased during the brace wearing period; however it returns to normal after completion of the brace wearing period . Nevertheless, the relationship between the compliance measure and the difference in TAPS mean score between Visit 2 and Visit 3 was shown not to be significant. This may have occurred because the patients perceived the effect of brace wear had already occurred by Visit 2 and thus it made no difference to their trunk perception by the time of Visit 3, regardless of their compliance.
The study results also predicted that the difference in BrQ mean score between Visit 2 and Visit 3 of the least compliant group (0–8 hours) would be significantly (−16.28) less than that of the most compliant group (17–23 hours). Moreover, a similar value was detected for the difference in the BrQ domain mean score, which included general health perception, physical functioning, emotional functioning as well as bodily pain. Thus, it appeared that the poorer compliance with brace wear, the poorer the QoL. This result contradicted the finding of Ugwonali et al. , who reported that brace wear did not decrease the QoL of adolescents with AIS. However, the questionnaire that Ugwonali et al. used was not condition-specific . Feise et al.  emphasised that disease-specific instruments are considered superior for measurements in homogenous populations because they concentrate primarily on the most significant domains of the disease and are more sensitive for measuring clinically important differences. In this study, the BrQ was adopted as a brace-oriented instrument. The questionnaire was translated into Chinese according to the guidelines of cross-cultural adaptation process used by the AAOS Outcomes Committee . It was then validated by the authors. The results showed that the Cronbach’s alpha and the intra-class correlation coefficient were both 0.93, and no floor or ceiling effect was demonstrated in all the BrQ domains except the bodily pain domain that showed a ceiling effect of 37.3%. Nevertheless, the result of this correlation study was in fact consistent with the finding of Rivett et al.  that poor compliance to a brace protocol is associated with poorer QoL. In the present study, in comparison to the most compliant subjects, the least compliant subjects tended to have a poorer perception of their general health and function more poorly both physically and emotionally. They also tended to have more bodily pain.
Regarding bracing hours, no significant difference was detected in the BrQ physical functioning and emotional functioning mean score difference between the moderately compliant group (9–16 hours) and the most compliant group (17–23 hours). However, in the most compliant group, the mean score difference tended to be less than that of the moderately compliant group. This is in line with the comment by Edgar , who reported 16 hours per day as the optimum wearing time to ensure a balance between the effectiveness of and tolerance to brace wear. In the study of Weinstein et al. , it also highlighted brace wear for an average of at least 12.9 hours per day was associated with success rates of 90 to 93%.
Regarding use of the orthosis monitoring system, it reliably generated objective data pertaining to patients’ compliance with bracing as shown in the present study. The mean wearing hours as recorded on the log sheets by the subjects was comparable to the mean wearing hours as read by the systems. No underreporting or overreporting of the wearing hours was observed. A significant correlation was also shown between the subjective compliance (wearing hours recorded on the log sheet) and the objective compliance (wearing hours recorded by the orthosis monitoring system). Hence, the system helps to prove if there is any overestimated duration of brace wear as recorded on the self-reporting inventory. Ultimately, such data gathering will help in establishing evidence-based AIS management.
Patient’s compliance with brace wear should not be undermined, as corrective bracing has shown favourable outcomes when the patient is compliant (3). Weinstein et al.  also supported that bracing significantly decreased the progression of high-risk curves to the threshold for surgery in patients with AIS. It also stressed longer hours of brace wear were associated with greater benefit as shown by the dose–response relationship. To enhance patient’s compliance, the orthotists at the local hospital adopted different strategies. One was making a hard copy of the immediate in-brace radiograph and presenting it to patient. The orthotists explained to the patient how the brace helped to control the curve progression. Upon each visit, checking on patient’s compliance was done by reviewing the wear and tear of the brace and the brace strapping, patient’s skill in wearing and removing the brace, and any skin discolouration on the pressure area of the patient as created by the brace. Ways to maximise the brace tolerability and reduce visibility were also introduced, such as changing to a new brace when the patient has grown significantly and making suggestions for wearing special clothing. In this study, the author, who had regular contact with the subjects, also checked the brace and the compliance records regularly. Otherwise, appropriate counselling would be provided to the patients after assessment of their emotional status. In fact, maintaining compliance is not viewed as the sole responsibility of the patients and their families. The responsibility also lies with the treating team, which may include orthopaedic surgeons, clinical psychologists, nurses, orthotists and physiotherapists. Their roles should be highlighted in terms of enhancing patient’s compliance. As recommended by the International Scientific Society on Scoliosis Orthopaedic and Rehabilitation Team (SOSORT), there is a need for a multiprofessional expert team to effectively treat the patients through increased compliance .
Quality of life measure
No relationship was found in the current study between in-brace correction and QoL. This differed from the finding of Vasiliadis et al. , who showed BrQ score to be related to the degree of deformity. The finding of present study suggested that QoL issues may be related more to psychosocial coping mechanisms than to physical deformity and its consequences . Freidee et al.  noted that patients with scoliosis also reported more physical complaints independent of the seriousness of the impairment, such as the magnitude of Cobb angle. Scoliosis causes psychological distress, regardless of the severity of curve . It is also true that successful treatment to prevent the progression of the curvature does not necessarily mean improvement in the QoL.
The present study showed that the brace affected QoL negatively and thus further reinforced the need to garner support from the multidisciplinary team for patients with AIS at the early stage of bracing. The ultimate purpose would be to manage patients’ psychosocial issues as provoked by the brace treatment. Reichel et al.  recognised that support for patients in the form of psychological group or individual sessions can help to prevent psychosocial impairment and it should be included in holistic management plans.
In exploring the relationship between self image and social functioning, a positive correlation was demonstrated by this study. As Deviren et al.  found, psychosocial and body image disturbances were less marked in patients with good social or family functioning. Again, it was shown to be important to provide support to patients and their families to enhance patients' QoL as well as patients' compliance with the treatment.
In this study, the BrQ and TAPS were shown to be effective in evaluating the QoL of patients with AIS. Both BrQ and TAPS exhibited superiority over the SRS-22r in detecting changes in QoL according to brace compliance. As Aulisa et al.  explained, this might be related to the greater number of questions contained in the BrQ that may allow it to explore more domains than the other questionnaires explore.