The results of this study indicate that compliance to a specific exercise programme and wearing the RSC brace can improve curvatures and signs and symptoms of AIS. The natural history of scoliosis was altered in the compliant subjects with non-compliance resulting in significant progression of the curvatures. Some studies have shown that conservative management of bracing and exercise has no effect on the natural history of scoliosis while others have shown that conservative management is effective [4, 23], but compliance was not reported on in these studies. The importance of compliance is supported by Landauer et al.  and Weinstein et al. , and should be monitored in future studies .
The risk of progression of the curvatures in both groups of subjects was high, 70-78 percent. The mean in-brace correction was 44.5% in the compliant group and 35.8% in the non -compliant group, which was not statistically different with similar in-brace corrections being reported [56, 57]. Compliance and initial correction effect in the brace are the two most important variables associated with good brace outcomes [5, 24, 25, 58]. The reason for the good in-brace correction in this study is probably the good three dimensional design of the RSC brace .
A study by Landauer et al.  has similar results to this study. They advised full time Cheneau bracing and weaned the subjects at Risser 5 over a six month period. They did not state how many hours the brace was actually worn, however, they did use a compliance score. The final overall outcome of their study, including all subjects, was that the thoracic curve improved three degrees only, not a successful result. However, when the subjects were divided into compliant and non compliant groups, then the study was very successful, as high early in-brace correction (40% or more) and good compliance achieved a correction of seven degrees in the Cobb angle. Low early correction (less than 40%) and good compliance resulted in stabilisation of the curvatures. Poor compliance with a high or a low initial correction resulted in progression of the Cobb angle  therefore emphasising the importance of evaluating compliance which is more important than in-brace correction.
The compliant group in this study had improved Cobb angles, angle of vertebral rotation and angle of trunk rotation (ATR), by wearing the brace 21.5 hours per day. Wearing the brace for 12 hours in the non- compliant group resulted in progression of the curvatures. Many Cheneau brace studies do not mention the number of hours the brace was worn [17, 54, 60, 61] but good brace compliance has been shown by a number of studies to have good outcomes [21–24, 62–65]. Recording the brace wearing hours was one of the challenges of this study. Using a compliometer or thermobrace, not currently used in South Africa, would have been more accurate in measuring actual brace wearing hours. Subjects’ diaries were checked frequently by the first author and parents for the record of brace hours. A study by Takemitsu et al.  showed that patients complied with 75 percent of a prescribed routine and on average over-reported their hours of brace wear to their physicians. The actual brace hours were measured using a compliance monitor. A study by Donzelli et al.  in which they used a temperature sensor (Thermobrace) showed compliance to be higher than previously reported with brace prescription being 16-23 hours per day and more than half the patients had 90 percent compliance.
The exercise programme was completed 3.9 times a week by the compliant group and 1.7 times a week by the non -compliant group This prescription of four to five times a week for 20-25 minutes is similar to the Schroth method , Dobosiewicz method , and SEAS and Side Shift method .
The Cobb angles were significantly reduced in the compliant group with the worst/major Cobb angle significantly improving (10.19 degrees), thus the natural history of AIS was altered. In the non -compliant group there was significant progression of the curvatures (5.5 degrees), following the natural history of scoliosis and three progressed to surgery. Supporting this study is a small study by Wood  on 23 subjects, with Cobb angles greater than 30 degrees, using the Cheneau brace over a four year period with a progression risk of 68 percent. Compliance was not measured in the Wood study nor was the physiotherapy described. Wood  showed that the major Cobb angle improved a mean of 13.2 degrees and the minor Cobb angle improved eight degrees at the end of brace wearing. Physiotherapy has been shown to have favourable outcomes in scoliosis patients and Rigo claims that physiotherapy can improve the actions of the Cheneau brace, by making the curve more flexible and preventing muscle atrophy [17, 70]. Cinnella et al.  using the Cheneau brace showed a 23 percent correction in Cobb angle at the end of a mean treatment period of 4.5 years, and after five years the correction was 15 percent but compliance was not monitored. Other studies on the Cheneau and RSC braces only show stabilisation of curvatures and reduction in children requiring surgery [6, 14, 64, 65, 72–74].
Kyphosis angles improved in the compliant group and deteriorated in the non- compliant group, and therefore the sagittal profile improved in the compliant group. The compliant group is similar to several studies that show normalisation of kyphosis and lordosis with the Cheneau brace [13, 17, 60]. In many Cheneau brace studies however, the sagittal profile is not mentioned, [24, 56, 59, 61, 65, 71]. The RSC brace allows sagittal normalisation because of its physiological profile in the sagittal plane and every trunk section is aligned to allow a normal sagittal profile [17, 75].
The angle of axial rotation predicts the incidence of progression of a curve better than the size of the Cobb angle . In this study all apical vertebra were significantly reduced in rotation in the compliant group p < 0.0001 in the worst/major curves which is similar to the study by Wood . The RSC and the Cheneau brace has also been shown to improve the wedge deformity of the apical vertebra, by over 50% in some cases [77, 78].
Scoliometer readings improved significantly in the compliant group. Similarly Kinel et al. [63, 79] showed that girls with AIS, wearing a Cheneau brace, revealed less clinical deformity than a group of non- treated girls with similar radiological deformities. Other studies using the Cheneau brace have shown cosmetic improvements in the deformity [13, 17, 59, 60, 80].
The peak flow changes in the compliant group improved significantly (by a mean of 20 percent) more than the non -compliant group (by nine percent). The Dobomed method has also been shown to improve exercise efficiency significantly using ergospirometry [69, 81]. Other studies using outpatient or in-patient Schroth method have shown significant improvements in vital capacity (VC) [82–85]. Dos Santos Alves et al.  showed that aerobic exercises three times a week for an hour, over a period of four months, resulted in a significant improvement in FVC, FEV1, inspiratory capacity, expiratory reserve volume and in respiratory muscle strength . Spirometry is the method of choice to identify any changes in the course of a respiratory disease  although a Mini Wright peak flow meter was used here to measure FEV1, as it is simple, portable, reproducible and practical to use clinically [89, 90].
Peak height was reached between 3.46 (±0.5) to 4.03(±0.6) years after menarchy. Weaning the subjects out of full time bracing, at the end of the study, occurred once height had been static for six months and as close to Risser 5 as possible. Weaning out the brace in the compliant group of this study started at the mean age of 17.14 (±0.6) years, which is later than most studies, in order to prevent progression, previously reported and postural collapse [12, 91]. The fact that peak bone mineralization and peak muscle strength occurs at 25, and peak ligamentous stability occurs in the early 20’s , was considered when deciding on weaning. Bracing is sometimes considered ineffective, when actually the subjects have just been weaned out too early or too quickly . The optimal weaning process is not known and has not been standardised [8, 12] and therefore was not included in this study. Skeletal maturity measures are not accurate enough to predict spinal growth potential in AIS .
One subject stopped wearing the brace in the non- compliant group. Psychosocial and body image disturbance are less marked in patients with good social and family functioning, as well as patients who exercise regularly  and this is similar to the compliant group in this study. The compliant group had larger Cobb angles than the non compliant group at baseline, therefor the severity of the scoliosis as measured by Cobb angle was not related to a poorer quality of life. Other Cheneau brace studies have shown the brace negatively affects quality of life [95–97]. Quality of life issues may be related to psychosocial coping mechanisms more than the physical deformity and its consequences. Support for AIS patients in group or individual sessions prevents psychosocial impairment, body image disturbances and should be included in holistic management plans [34, 40]. Programmes to address personal, group and family issues may improve QoL, promoting compliance . In this study emotional function was lower in the non- compliant group, and subjects did not believe that the brace was beneficial, had low self esteem and low social function. Lindeman and Behm , showed that non- compliant girls did not expect to succeed in dealing with scoliosis, they were anxious about possible failure, had low self- esteem and did not seek social support.
Non compliance to the intervention resulted in progression in curvatures. These subjects had a poorer quality of life and seemed to need psychosocial support to improve compliance and therefore treatment outcomes. This should be included as part of a management programme, as subjects do not generally seek help. Continuous monitoring of stress and QoL needs to be done, which will allow modification of the treatment and maintain good compliance  with regular consultations with a psychologist and family counselling.
The personality trait questionnaire revealed, that the compliant group was more emotionally mature, stable and realistic than the non- compliant group, which was more emotionally immature and assertive (Factor C). Higher scores in Factor C (compliant group) reveal emotional stability, control, and high ego strength . These patients appear calm, unruffled, behave in an adult and rational way, they are realistic, constant in interests, responsible, distinguish between emotional needs and reality, and adjust to facts. High Factor C scores also correlate with positive family relationships and leadership . Lower Factor C scores, in the non- compliant group, indicated emotional instability and low ego strength . They reveal an inability to control their emotions, impulses and to find satisfying and realistic ways of expressing them. They are easily angered, are more frequently dissatisfied with their family and school, find it difficult to restrain themselves and are discouraged by their inability to meet good standards of behaviour. They are easily perturbed, confused, changeable in attitudes and interests. They evade responsibility, give up easily, tend to worry a lot, have irrational fears and get into fights and problem situations. They can experience severe adjustment problems if subjected to regimentation and stress  with bracing being stressful . These factors may explain their lack of compliance.
Psychological support during this adjustment to bracing and exercise phase as well as during the treatment therefore seems to be essential. The results of the personality questionnaire compare well with the results of the BQ, which showed that the non- compliant subjects had a poorer QoL.
Other traits from the questionnaire, revealed marginally significant differences in that the compliant group was more obedient, mild and dependant (Factor E), tender minded, sensitive, protected (Factor I) relaxed, and composed (Factor Q4). The non- compliant group was more aggressive, rebellious, dominant (Factor E), practical tough- minded and tense (Factor I) driven, irritable, and frustrated (Factor Q4). In Factor E, lower scores, in the compliant group, show traits that are more accommodating, more compliant and easily influenced . High scores in Factor E seen in the non- compliant group, show traits of stubbornness, and are headstrong, arrogant and disobedient.
This study has determined the personality traits of compliant and non- compliant subjects and these traits can be used to predict compliance of a subject. Should a patient be predicted to be non -compliant using the HSPQ, at the beginning of treatment, then appropriate interventions, such as regular individual and family counselling could be implemented at the beginning of the programme in an attempt to improve compliance. The up to date personality questionnaire now recommended by the editors of the HSPQ is the “16PF Adolescent Personality Questionnaire”, by Scheurger , for 11-22 year olds.