SPINE
An alternative to a randomised control design for assessing the efficacy and effectiveness of bracing in adolescent idiopathic scoliosis D. Y. T. Fong, K. M. C. Cheung, Y. W. Wong, W. Y. Cheung, I. C. Y. Fu, E.E. Kuong, K. C. Mak, M. To, D. Samartzis, K. D. K. Luk From The University of Hong Kong, Hong Kong, China D. Y. T. Fong, PhD, Associate Professor, School of Nursing I. C. Y. Fu, DN, Lecturer, School of Nursing The University of Hong Kong, 21 Sassoon Road, Pokfulam, Hong Kong SAR, China. K. M. C. Cheung, MBBS, MD, FRCS, FHKCOS, FHKAM (Orth), Clinical Professor and Head, Department of Orthopaedics & Traumatology Y. W. Wong, MBBS, FRCSE, FHKCOS, FHKAM (Orth), Consultant, Department of Orthopaedics & Traumatology W. Y. Cheung, MBBS, FRCSE, FHKAM, FHKCOS, Clinical Assistant Professor, Department of Orthopaedics & Traumatology E.E. Kuong, MBBS, Associate Consultant, Department of Orthopaedics & Traumatology K. C. Mak, MBBS, Associate Consultant, Department of Orthopaedics & Traumatology M. To, MBBS, FRCSEd, FHKCOS, FHKAM, Clinical Assistant Professor, Department of Orthopaedics & Traumatology D. Samartzis, DSc, Assistant Professor, Department of Orthopaedics & Traumatology K. D. K. Luk, MBBS, MChOrth, FRCSEd, FRCSGlas, FRACS, FHKAM, Chair Professor, Department of Orthopaedics & Traumatology The University of Hong Kong, 5/F, Professorial Block, Queen Mary Hospital, Pokfulam, Hong Kong SAR, China. Correspondence should be sent to Dr K. D. K. Luk; e-mail: [email protected] ©2015 The British Editorial Society of Bone & Joint Surgery doi:10.1302/0301-620X.97B7. 35147 $2.00 Bone Joint J 2015;97-B:973–81. Received 18 September 2014; Accepted after revision 5 March 2015
VOL. 97-B, No. 7, JULY 2015
Randomised controlled trials (RCTs) that assessed the efficacy of bracing for adolescent idiopathic scoliosis have suffered from small sample sizes, low compliance and lack of willingness to participate. The aim of this study was to assess the feasibility of a comprehensive cohort study for evaluating both the efficacy and the effectiveness of bracing in patients with adolescent idiopathic scoliosis. Patients with curves at greater risk of progression were invited to join a randomised controlled trial. Those who declined were given the option to remain in the study and to choose whether they wished to be braced or observed. Of 87 eligible patients (5 boys and 63 girls) identified over one year, 68 (78%) with mean age of 12.5 years (10 to 15) consented to participate, with a mean follow-up of 168 weeks (0 to 290). Of these, 19 (28%) accepted randomisation. Of those who declined randomisation, 18 (37%) chose a brace. Patients who were more satisfied with their image were more likely to choose bracing (Odds Ratio 4.1; 95% confidence interval 1.1 to 15.0; p = 0.035). This comprehensive cohort study design facilitates the assessment of both efficacy and effectiveness of bracing in patients with adolescent idiopathic scoliosis, which is not feasible in a conventional randomised controlled trial. Cite this article: Bone Joint J 2015; 97-B:973–81
Bracing has a long history in the treatment of adolescent idiopathic scoliosis (AIS), and it is recognised that patients who present radiographically with coronal curves are at high risk of progression.1 However, reports are mixed on whether bracing reduces curve progression and the need for surgery.2 A prospective cohort study by the Scoliosis Research Society (SRS) reported that bracing could prevent curve progression more effectively than observation alone.3 However, this study was not randomised with concurrent controls. In addition, there have been five systematic reviews or meta-analyses of bracing, including one underpowered randomised controlled trial (RCT).2,4-8 There have been only two large multicentre RCTs that have assessed the efficacy of bracing for AIS,9,10 but both encountered recruitment difficulties.9,11 Although an RCT design is regarded as the ideal for establishing efficacy, it has limitations, especially when used to assess the effectiveness of bracing in practice. This is because no-one can be blinded to bracing, and their preferences and expectations may influence the effectiveness of of this treatment.12,13 Cultural, social and economic factors may also affect decisions by both patients and parents to participate in an RCT.14 There have been reports
of poor participation in RCTs, which severely limit the generalisability of study findings.9 Therefore, it is important to find out what both patients and their families think of the available therapeutic options for AIS, which cannot be addressed in a traditional RCT design.15 We have proposed an alternative study design to evaluate both the efficacy and the effectiveness of bracing over observation. In order to do this we conducted a feasibility study to obtain preliminary information.
Patients and Methods We designed a comprehensive cohort study (CCS) in which treatment with bracing or observation was determined either by randomisation or by the preference of the patients and/ or the parents. Eligible patients were followedup every four months for a minimum of 12 months. Ethical approval was granted by a local institutional research board (NCT00989495). The CCS design is a new paradigm of evidencebased clinical research that combines an RCT and a preference-based controlled trial.12 The treatment allocation plan is shown in Figure 1. The trial examines the treatment efficacy, as in a traditional RCT (Group I vs Group II in Fig. 1); it also addresses the issue of effectiveness, by 973
974
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
Assessed for eligibility by reading patient records (n = 6479)
Assessed for eligibility by orthopaedic surgeons (n = 423) Not meeting inclusion criteria (n = 336) Eligibility confirmed (n = 87) Declined to participate (n = 19) Yes
Group I:
Group II:
(Brace) (n = 13) • Braced (n = 11) • Did not brace (n = 2; withdrawn)
(Observation) (n = 6) • Observed (n = 6)
• Followed (n = 11) • Withdrawn (n = 2); no reason)
Visit 2
Visit 3
Baseline
Visit 1
No
Consent to randomisation?
Group III: Brace (Brace) (n = 9) • Braced (n = 9)
Group IV: Observation (Observation) (n = 19) • Observed (n = 18) • Braced (n = 1); made in Nanjing by parents)
Group V: Brace (Observation) (n = 9) • Braced (n = 9)
Group VI: Observation (Brace) (n = 12) • Observed (n = 12)
• Followed (n = 6) • Withdrawn (n = 0)
• Followed (n = 9) • Withdrawn (n = 0)
• Followed (n = 18) • Withdrawn (n = 1); no reason)
• Followed (n = 9) • Withdrawn (n = 1); mother refused to complete questionnaires)
• Followed (n = 12) • Withdrawn (n = 0)
• Followed (n = 11) • Withdrawn (n = 2)
• Followed (n = 6) • Withdrawn (n = 0)
• Followed (n = 8) • Withdrawn (n = 1); no reason)
• Followed (n = 18) • Withdrawn (n = 1)
• Followed (n = 9) • Withdrawn (n = 2); father refused to complete questionnaires)
• Followed (n = 12) • Withdrawn (n = 0)
• Followed (n = 11) • Withdrawn (n = 2)
• Followed (n = 6) • Withdrawn (n = 0)
• Followed (n = 8) • Withdrawn (n = 1)
• Followed (n = 18) • Withdrawn (n = 1)
• Followed (n = 9) • Withdrawn (n = 2)
• Followed (n = 12) • Withdrawn (n = 0)
Fig. 1 Flow chart of patient disposition. For patients who consented to be randomised, the treatment in brackets indicates what was decided by the randomisation. For patients who declined randomisation and chose their treatments, the treatments in brackets indicate what would have been given if randomised, which may not necessarily be the same as the treatments they actually received.
assessing the impact of the treatment preference (Group I vs Group II and Groups III + V vs Groups IV + VI in Fig. 1). Because of the concurrent assessment of efficacy and effectiveness, the inclusion of non-randomised subjects has even been suggested in all RCTs to assess external validity.16 Patients aged ten years or older with Risser17 sign 0 to 2, and diagnosed with AIS with either a Cobb18 angle of 20° to < 25° and at least 5° deterioration over the past four months, or a Cobb angle of 25° to < 30°, were invited to participate. Exclusion criteria were patients with a history of treatment for AIS; who were more than year postmenarche; who had a physical or mental disability that interfered with adherence to bracing; who had a diagnosed musculoskeletal or developmental illness that might be responsible for the spinal curvature; who had a structural thoracic scoliosis with the apex above T7; who had difficulty in understanding or completing the study questionnaires; or who were unlikely to comply with the study protocol. These criteria conformed with those suggested by the SRS for bracing studies.19 We planned to recruit at least
ten patients into each of the two RCT groups and at least five into each of the other four groups (Fig. 1). Hence, a total of at least 40 patients were to be recruited for the pilot study.20 Interventions. For patients who underwent bracing, a custom-made Boston-type of thoracolumbar sacral orthosis (TLSO)21 was moulded under traction by an orthotist. The brace extended anteriorly from the pubis to the xiphoid process. Patients were asked to wear the brace for a minimum of 20 hours per day, and the brace straps were marked to ensure that correct pressure was applied. Patients under observation only returned for assessment as scheduled. Study procedures. Before patient recruitment, a treatment allocation sequence was generated by means of block randomisation, with block size randomised as four or eight. The allocation sequence was concealed using opaque envelopes, each containing the allocation code of one patient. Patients were recruited from the clinic of a children’s hospital largely devoted to orthopaedic patients. Before each clinic session, a recruitment team composed of a THE BONE & JOINT JOURNAL
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
975
Table I. The sociodemographic characteristics of the 68 study patients and their parents
Mean age (yrs) (standard deviation SD) Gender Male (n, %) Female (n, %) Educational level Grade 5 to 6 (n, %) Grade 7 to 11 (n, %) Mean height (cm) (SD) Mean weight* (kg) (SD) Mean BMI* (kg/m2) (SD) Mean arm span (cm) (SD) Mean monthly household income (HK$) < 2000 (n, %) 2000 to 5999 (n, %) 600 to 9999 (n, %) 10 000 to 14 999 (n, %) 15 000 to 19 999 (n, %) 20 000 to 24 999 (n, %) 25 000 to 29 999 (n, %) 30 000 to 39 999 (n, %) 40 000 to 59 999 (n, %) ≥ 60 000 (n, %) Mean gestational age (wks) (SD) Mean maternal age (yrs) (SD) Breastfeeding No (n, %) Yes (n, %) Do not know/Missing (n, %) Smoking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Alcohol drinking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Family history of scoliosis No (n, %) Yes (n, %) Do not know/missing (n, %)
RCT component (n = 19)
Preference component (n = 49)
p-value
12.4 (1.1)
12.6 (1.1)
0.383
1 (5.3) 18 (94.7)
4 (8.2) 45 (91.8)
0.684
7 (36.8) 12 (63.2) 150.9 (8.0) 38.3 (7.9) 16.7 (2.6) 152.2 (10.2)
18 (36.7) 31 (63.3) 154.8 (5.6) 41.9 (8.6) 17.4 (3.1) 155.5 (7.0)
2 (10.5) 1 (5.3) 2 (10.5) 3 (15.8) 4 (21.1) 3 (15.8) 0 (0) 1 (5.3) 1 (5.3) 2 (10.5) 39.7 (1.4) 30.3 (4.3)
1 (2.0) 2 (4.1) 5 (10.2) 3 (6.1) 11 (22.4) 6 (12.2) 3 (6.1) 5 (10.2) 5 (10.2) 8 (16.3) 39.4 (1.2) 30.0 (5.0)
13 (68.4) 6 (31.6) 0 (0)
34 (69.4) 13 (26.5) 2 (4.1)
13 (68.4) 6 (31.6) 0 (0)
27 (55.1) 20 (40.8) 2 (4.1)
12 (63.2) 7 (36.8) 0 (0)
26 (59.2) 20 (40.8) 3 (6.1)
14 (73.7) 1 (5.3) 4 (21.1)
29 (59.2) 8 (16.3) 12 (24.5)
0.993
0.039 0.125 0.404 0.138 0.096
0.393 0.805 0.750
0.411
0.622
0.223
* Three missing standard deviation; BMI, body mass index
SD,
research assistant and two of the authors (DF, IF) screened all the clinical records and identified potential patients. These patients were seen by orthopaedic surgeons to assess their eligibility. Patients who were suitable were referred to the recruitment team, who explained the study details and sought informed consent. Each eligible patient was invited to participate in an RCT. Those who declined were given an option to stay in the study and choose whether they wished to be braced or observed. The consent form was signed together with one of their accompanying guardians or parents. The 60 patients with the best study compliance received about $25 (USD) to offset their travel expenses. Braced patients were telephoned at least one day before their next clinical visit to remind them to remove the brace at least 24 hours before their visit. In the literature, the length of time out-of-brace varies, but at least 24 hours free VOL. 97-B, No. 7, JULY 2015
from the brace would ensure full relaxation of the spine for a more realistic measurement of the true curvature.22,23 Measurements. The sociodemographic details of all patients were recorded. At all clinical visits, height (cm), weight (kg), arm span (cm), menarche status and exercise habits were measured. All participants self-completed the Scoliosis Research Society-22 (SRS-22) questionnaire, the State-Trait Anxiety Inventory (STAI) and the Beck Depression Inventory version II (BDI-II).24-26 These tools have been shown to have good validity and reliability in Chinese subjects.27-29 In addition, standing posteroanterior out-ofbrace radiographs of the full spine were taken. An orthopaedic surgeon (EK) performed blinded assessment of the radiographs. Only one observer was used for assessment of the radiographs to avoid interobserver differences, as recommend by the SRS.19 Braced patients were also given a
976
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
Table II. The curvature and quality of life measures of the patients and their parents RCT Component (n = 19)
Cobb angle (º) State anxiety Patient Mother Father Trait anxiety Patient Mother Father Beck Depression Inventory Patient Mother Father SRS-22 Function Pain Mage Mental health Management Subtotal Total
Preference component (n = 49)
n
Mean (SD)
n
Mean (SD)
p-value
18
25.4 (3.5)
48
25.5 (3.5)
0.938
19 16 4
46.0 (11.2) 48.9 (8.2) 45.3 (7.2)
49 38 15
43.9 (8.9) 49.8 (12.8) 44.7 (8.3)
0.404 0.790 0.905
19 16 4
42.2 (7.8) 45.4 (5.8) 42.3 (5.4)
49 38 15
41.7 (8.5) 45.0 (7.5) 39.4 (7.4)
0.826 0.845 0.462
19 16 4
5.6 (6.6) 4.1 (4.2) 6.3 (1.0)
48 36 15
4.5 (4.6) 6.6 (6.5) 3.3 (5.1)
0.420 0.173 0.269
18 18 18 18 15 18 15
4.9 (0.1) 4.7 (0.4) 3.8 (0.6) 3.8 (0.6) 3.8 (0.6) 4.4 (0.3) 4.3 (0.3)
43 43 43 43 34 43 34
4.9 (0.2) 4.7 (0.4) 3.9 (0.6) 4.1 (0.5) 3.5 (0.6) 4.4 (0.3) 4.3 (0.3)
0.635 0.968 0.781 0.435 0.130 0.711 0.707
RCT, Randomised controlled trial; standard deviation; SRS, Scoliosis Research Society
SD,
diary to record the daily number of hours in the brace and asked to return the completed diaries at each follow-up visit. An accompanying parent of each patient also selfcompleted the STAI and BDI-II. Statistical analysis. Quality of life measures were scored in accordance with the corresponding scoring manuals.25,26 Specifically, the SRS-22 has domains scored as means in the range of 1 to 5. If fewer than half of the items in a domain had responses, the skipped items were scored by the mean of the other items in the domain. Otherwise, the domain score was recorded as missing. Determinants of acceptance to randomisation were examined by comparing groups I + II and III + IV + V + VI (described in Fig. 1) by a stepwise logistic regression on all baseline measurements. The effects of bracing on quality of life and SRS-22 were assessed by a linear mixed effects model to account for repeated measurements over time, whereas those at risk of ≥ 6º progression were assessed by Cox’s regression analysis. All significance tests were two-tailed, and a p-value < 0.05 was deemed significant. Analysis was performed using SAS v 9.2 software (SAS Institute, Cary, North Carolina).
Results Between 9 March 2009 and 8 March 2010, a total of 87 eligible AIS patients were approached and 68 (78%) consented to participate in the study. About six patients per month were recruited. Among the 19 refusals, 16 (84%) were made by the parents (11 mothers and five fathers) and three were made by the children (Fig. 1). Parents refused to participate because they were unwilling to be involved in a
research study (n = 7); they wanted to decide at the next visit (n = 6); they were in a hurry (n = 2); or they thought a retrospective study was possible without consent (n = 1). One newly diagnosed girl was too emotionally unstable to participate. Another girl whose mother consented to randomisation was afraid of bracing and refused to participate. The other child (a boy) refused participation because he was unwilling to complete the questionnaires. Of the 68 study patients (five boys, 63 girls), 19 (28%) consented to be randomised (Tables I and II). The participation rate in the RCT was 19/87 (22%). Among the 49 patients who chose their own treatment, 18 (37%) chose bracing (Tables III and IV). Taller patients (Odds Ratio (OR) 0.90, 95% CI 0.81 to 0.995; p = 0.040) or those who had a higher monthly household income (OR 0.75, 95% CI 0.57 to 0.98; p = 0.032) preferred to choose their treatment (Table V). Among patients who had their treatment preferences, those with one unit score higher in the domain ‘satisfaction with self-image’ were three-times more likely to choose bracing (OR 4.1, 95% CI 1.1 to 15.0; p = 0.035) (Table V). Patients were followed for a maximum of one year on the self-reported assessments, but the follow-up of their curvature measurements was extended to January 2015. The mean duration of follow-up was 168 weeks (0 to 290), and six patients withdrew from the study within the first year of follow-up. The parents of two patients refused to continue with the study because of unwillingness to complete the questionnaires. Another three did not continue follow-up at the specialist hospital after recruitment, and never reattended. The other two patients did not provide reasons THE BONE & JOINT JOURNAL
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
977
Table III. The sociodemographic characteristics of patients who desired a certain treatment
Mean age (yrs) (SD) Gender Male (n, %) Female (n, %) Educational level Grade 5 to 6 (n, %) Grade 7 to 11 (n, %) Mean height (cm) (SD) Mean weight* (kg) (SD) Mean BMI* (kg/m2) (SD) Mean arm span (cm) (SD) Monthly household income (HK$) < 2000 (n, %) 2000 to 5999 (n, %) 6000 to 9999 (n, %) 10 000 to 14 999 (n, %) 15 000 to 19 999 (n, %) 20 000 to 24 999 (n, %) 25 000 to 29 999 (n, %) 30 000 to 39 999 (n, %) 40 000 to 59 999 (n, %) ≥ 60 000 (n, %) Mean gestational age (wks) (SD) Mean maternal age (yrs) (SD) Breastfeeding No (n, %) Yes (n, %) Do not know/missing (n, %) Smoking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Alcohol drinking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Family history of scoliosis No (n, %) Yes (n, %) Do not know/missing (n, %)
Brace (n = 18)
Observation (n = 31)
p-value
12.4 (0.9)
12.7 (1.2)
0.275 0.616
1 (5.6) 17 (94.4)
3 (9.7) 28 (90.3)
7 (38.9) 11 (61.1) 153.7 (5.4) 41.0 (8.7) 17.3 (2.9) 154.6 (7.1)
11 (35.5) 20 (64.5) 155.4 (5.8) 42.5 (8.6) 17.5 (3.3) 156.1 (7.1)
1 (5.6) 0 (0) 3 (16.7) 1 (5.6) 1 (5.6) 3 (16.7) 2 (11.1) 3 (16.7) 2 (11.1) 2 (11.1) 39.7 (1.3) 28.2 (5.1)
0 (0) 2 (6.5) 2 (6.5) 2 (6.5) 10 (32.3) 3 (9.7) 1 (3.2) 2 (6.5) 3 (9.7) 6 (19.4) 39.3 (1.1) 31.1 (4.7)
13 (72.2) 5 (27.8) 0 (0)
21 (67.7) 8 (25.8) 2 (6.5)
8 (44.4) 10 (55.6) 0 (0)
19 (61.3) 10 (32.3) 2 (6.5)
9 (50.0) 8 (44.4) 1 (5.6)
17 (54.8) 12 (38.7) 2 (6.5)
11 (61.1) 3 (16.7) 4 (22.2)
18 (58.1) 5 (16.1) 8 (25.8)
0.812
0.311 0.567 0.802 0.478 0.953
0.267 0.056 0.989
0.159
0.708
0.982
* Two missing SD, standard deviation; BMI, body mass index; HK$, Hong Kong dollars
for their withdrawal. Of 29 braced patients, 27 (93%) returned their diaries. The overall mean number of hours in the brace during the first year was 17.0 per day (95% CI 15.5 to 18.6), which was not significantly different between those randomised and those who chose to be braced (p = 0.634, by t-test). In the RCT arm there were no significant differences in anxiety and depression scores between braced and the observed patients (p > 0.212, by linear mixed effects model) (Fig. 2). In the preference arm no significant differences were found (p > 0.202, by linear mixed effects model), except that mothers of the braced patients had their depression level reduced significantly more than those of the observed patients (p = 0.031, by linear mixed effects model). Figure 3 similarly assesses the effects of bracing on the SRS-22. In both the RCT and the preference arms VOL. 97-B, No. 7, JULY 2015
braced patients scored significantly lower in the function, image, subtotal and total scales (p ≤ 0.042, by linear mixed effects model). In addition, braced patients in the preference arm also had significantly lower mental health scores (p = 0.030, by linear mixed effects model). A total of 33 patients had ≥ 6º of curve progression during follow-up, eight of whom (five braced and three observed) had treatment by randomisation and 25 (eight braced and 17 observed) chose their treatment. In the RCT arm the relative risk (RR) of curve progression of ≥ 6 º with bracing was 1.43 (95% CI 0.27 to 7.61, p = 0.679). In the preference arm the corresponding RR was 0.27 (95% CI 0.10 to 0.75, p = 0.012). There was a trend towards a more profound effect of bracing in the preference arm than in the RCT arm (ratio of RRs = 0.20, 95% CI 0.03 to 1.33; p = 0.095).
978
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
Table IV. The curvature and quality of life measures of patients who desired a certain treatment Brace (n = 18)
Cobb angle (º) State anxiety Patient Mother Father Trait anxiety Patient Mother Father Beck depression inventory Patient Mother Father SRS-22 Function Pain Image Mental health Management Subtotal Total SD,
Observation (n = 31)
n
Mean (SD)
n
Mean (SD)
p-value
18
26.4 (4.1)
30
24.9 (3.0)
0.160
18 15 4
44.9 (10.2) 31 53.5 (13.1) 23 37.5 (12.5) 11
43.3 (8.1) 47.5 (12.2) 47.4 (4.7)
0.532 0.161 0.141
18 15 4
41.6 (9.1) 31 45.33 (8.4) 23 37.5 (9.3) 11
41.7 (8.2) 44.7 (7.0) 40.1 (7.0)
0.968 0.808 0.539
17 15 4
4.8 (S4.8) 9.0 (8.3) 2.5 (3.8)
31 21 11
4.3 (4.5) 4.9 (4.3) 3.6 (5.6)
0.716 0.079 0.717
15 15 15 15 10 15 10
5.0 (0.1) 4.7 (0.3) 4.1 (0.5) 4.2 (0.6) 3.8 (0.7) 4.5 (0.3) 4.4 (0.3)
28 28 28 28 24 28 24
4.9 (0.2) 4.7 (0.5) 3.7 (0.6) 4.1 (0.5) 3.5 (0.6) 4.4 (0.3) 4.3 (0.3)
0.259 0.712 0.035 0.400 0.213 0.112 0.232
standard deviation; SRS, Scoliosis Research Society
Table V. Determinants of acceptance of randomisation and choosing brace treatment, by stepwise logistic regression
Determinants of acceptance of randomisation (n = 68) Patient’s height (cm) Monthly household income (HK$) Determinants of choosing brace treatment (n = 49) Satisfaction with self-image
Odds ratio
95% CI
p-value
0.90 0.75
(0.81 to 0.995) (0.57 to 0.98)
0.040 0.032
4.1
(1.10 to 15.0)
0.035
CI, confidence interval
RCT arm absolute p-values: STAI_state (patient), p = 0.645; STAI_trait (patient), p = 0.212; BDI (patient), p = 0.583; STAI_state (mother), p = 0.719; STAI_trait (mother), p = 0.272; BDI (mother), p = 0.645. Preference arm absolute p-values: STAI_state (patient), p = 0.293; STAI_trait (patient), p = 0.202; BDI (patient), p = 0.229; STAI_state (mother), p = 0.980; STAI_trait (mother), p = 0.576; BDI (mother), p = 0.031. RCT arm absolute p-values: Function, p = 0.042; Image, p = 0.035; Subtotal, p = 0.011; Total, p = 0.018. Preference arm absolute p-values: Function, p = 0.001; Image, p < .001; Subtotal, p < 0.001; Total, p = 0.002.
Discussion The CCS design addresses the concerns raised by traditional RCTs for bracing, owing to the low acceptance of treatment decided by randomisation. The participation rate in our RCT component was 21%, the same as that in a Dutch RCT but lower than the 33% reported in a hypothetical setting.9,11 Refusal to consent to enter an RCT as a result of patients’ preference as to their treatment was
observed in this study. The effect of bracing on the risk of ≥ 6º progression in the preference arm was larger than that in the RCT arm, although the difference did not reach statistical significance owing to the small sample size. The potential influence of such preference on brace effects can be addressed by including the preference component under our CCS, which is what a traditional RCT cannot accomplish. The analysis of a CCS should begin with an assessment of efficacy of bracing by comparing the brace (Group I) and the observation (Group II) groups under the RCT component. Then the demographic and socioeconomic status of patients between the RCT (Groups I + II) and the preference (Groups III + IV + V + VI) components are compared with identify determinants of treatment preference. Finally, the impact of treatment preference on bracing effects is examined by assessing the preference by treatment interaction, i.e. (Groups I vs Group II) and (Groups III + V vs Groups IV + VI).30 When there is no indication of an influence of treatment preference, a more reliable estimate of bracing effects based on patients in the two components can be obtained. THE BONE & JOINT JOURNAL
50 40
60 50 40 30
20
20
BDI (patients)
50 40 30 20
SRS_Function
4
4
3 Observation Brace
Observation Brace
50 40 30
4 3
20
Observation Brace
2
60 50 40
70 60 50 40
70 60 50 40 30
1 5
4
4
3 Observation Brace
Observation Brace
70 60 50 40
20 60
Observation Brace
Observation Brace
40 30 20
SRS_subtotal -
50 BDI (mother)
50
40 30 20
4 3 Observation Brace
2
2
3
0
Visit
1
2
3
Visit
Fig. 2 Graphs showing the longitudinal change in mean anxiety and depression scores by treatment group.
SRS_Total
1
5
4 3 Observation Brace
VOL. 97-B, No. 7, JULY 2015
Observation Brace
2 1
4 3 Observation Brace
2
1 5
1 5
4
4
3 Observation Brace
2
The treatment decision was mostly made by the patients’ parents, and most chose observation. Parents of taller patients think their children have little potential to grow taller and so they were more determined to choose observation. On the other hand, parents with greater financial resources were often more educated and had a better understanding of the risks and benefits of bracing. They were
3
5
SRS_Total
0 0
Observation Brace
4
1
10
0
3
1 5
2 10
Observation Brace
2
1 5
20 80
30
20 60
3
1 5
2
SRS_Management
STAI_Trait (mothers)
Observation Brace
4
Observation Brace
30
20 80
Observation Brace
2
SRS_MH
70
0 80
SRS_MH
Observation Brace
3
1 5
10
0 80
Observation Brace
2
1 5
60
Observation Brace
1 5
2
30
3
1 5
SRS_Pain
60
Observation Brace
70
4
2
Observation Brace
SRS_Image
Observation Brace
3
SRS_Management
20 80
70
4
2
20 80
STAI_State (mothers)
STAI_State (mothers)
40 30
30
STAI_Trait (mothers)
50
30
10
BDI (mother)
60
979
5
SRS_subtotal -
40
SRS_Function
50
5
Observation Brace
70
SRS_Pain
60
80
SRS_Image
Observation Brace
70
60 BDI (patients)
STAI_State (patients)
80
STAI_Trait (patients)
STAI_Trait (patients)
STAI_State (patients)
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
3 Observation Brace
2
1
1 0
1
2
3
0
Visit
1
2
3
Visit
Fig. 3 Graphs showing the longitudinal change in mean SRS-22 component scores by treatment group.
more prone to express preferences and to refuse allocation of treatment by randomisation.
980
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
It is interesting to note that greater patient satisfaction with their image could influence their parents to choose bracing. A main consideration of parents was whether their children would accept bracing, which has previously been shown to reduce the satisfaction of patients with their image.31 It could be that patients who had greater satisfaction with their image have greater tolerance of their appearance and thus are more open to bracing. On the other hand, the satisfaction level with image in patients who chose observation, albeit lower than that in patients who chose bracing, was still good. These patients did not exhibit significant dissatisfaction with their curvature, to the extent that it would make them choose to be braced to improve their appearance. Several important aspects must be considered when planning for the full-scale CCS. First, during patient recruitment, it is important that all orthopaedic surgeons and patient recruiters remain impartial when offering advice to patients. For example, it has been the preference of the surgeons in our clinic to brace skeletally immature patients with curves ≥ 30° because of the perceived higher risk of progression.32 These surgeons have found it difficult to stay impartial when offering advice to the group in this study. Hence it was decided to restrict the inclusion criterion to curves that < 30°, so that the recruiters could comfortably explain that there is no good evidence available to support either bracing or observation. Second, there were unbalanced numbers of patients across the different treatment groups. Balanced group sizes would result in smaller standard errors and hence give larger power to detect an effect. Therefore, block randomisation stratified by whether patients consent to randomisation or not is recommended. Third, it is preferable to allow full relaxation of the spine to facilitate a realistic measure of the curvature. Fourth, temperature sensors could give a more accurate record of the actual hours of brace wear,33 However, there is still no guarantee that the strap tensions were appropriately tightened. In fact, pressure sensors have been developed to monitor both the duration and the quality of bracing.34 Fifth, we did not find that the financial incentive added to the participation of parents in our study, as parents were predominantly more concerned about which was the best treatment for their children, rather than any monetary compensation. Sixth, several parents were unwilling to complete the questionnaires covering the psychosocial domains. Strategies would be formulated to enhance participants’ compliance. A study by Weinstein et al10 with a similar design, which began recruiting subjects after our initial study recruitment, was terminated prematurely owing to the effect of bracing over observation in reducing curve progression. As in our study, Weinstein et al had difficulty in recruiting patients for the RCT arm of their study, and they noted the inclination of patients to choose the treatment type option. However, their study population was based on a heterogeneous
Western population and not an Asian one, which may also present with specific ethnic and cultural considerations related to study and treatment selection, such as degree of cosmetic concern and risk of skin irritation, and hence affect the outcomes. Our study, in which it was impossible to blind patients, parents and physicians to brace treatment, has revealed that parents have a strong preference to choose their desired treatment, which limits the role of traditional RCTs. We have successfully completed a feasibility study of a CCS design that allows the assessment of the impact of preference on the effectiveness of bracing. A full-scale trial is now being planned, which can be further used to address the efficacy and effectiveness between different populations. Author contributions: D. Y. T. Fong: Research design, Case recruitment, Data collection, Analysis, Writing paper. K. M. C. Cheung: Research design, Case recruitment, Data collection, Analysis, Writing the paper. Y. W. Wong: Research design, Case recruitment, Data collection, Analysis, Writing the paper. W. Y. Cheung: Case recruitment, Data collection, Writing the paper. I. C. Y. Fu: Case recruitment, Data collection, Writing the paper. E. E Kuong: Case recruitment, Data collection, Writing the paper. K. C. Mak: Case recruitment, Data collection, Writing the paper. M. To: Case recruitment, Data collection, Writing the paper. D. Samartzis: Analysis, Writing the paper. K. D. K. Luk: Research design, Case recruitment, Data collection, Analysis, Writing the paper. The work was financially supported by the Scoliosis Research Society. We thank Sister Leung, Ms J. Lam, Ms SY Lam and all nursing staff at the Duchess of Kent Children’s Hospital for facilitating the subject recruitment. We also thank Ms NL Kwok for recruiting the patients and coordinating the study, as well as Ms P. Wan and Mr G. Chan for following up the patients. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This article was primary edited by S. P. F. Hughes and first proof edited by G. Scott.
References 1. Asher MA, Burton DC. Adolescent idiopathic scoliosis: natural history and long term treatment effects. Scoliosis 2006;1:2. 2. Maruyama T, Grivas TB, Kaspiris A. Effectiveness and outcomes of brace treatment: a systematic review. Physiother Theory Pract 2011;27:26–42. 3. Nachemson AL, Peterson LE. Effectiveness of treatment with a brace in girls who have adolescent idiopathic scoliosis. A prospective, controlled study based on data from the Brace Study of the Scoliosis Research Society. J Bone Joint Surg [Am] 1995;77-A:815–822. 4. Rowe DE, Bernstein SM, Riddick MF, et al. A meta-analysis of the efficacy of non-operative treatments for idiopathic scoliosis. J Bone Joint Surg [Am] 1997;79A:664–674. 5. Dolan LA, Weinstein SL. Surgical rates after observation and bracing for adolescent idiopathic scoliosis: an evidence-based review. Spine (Phila Pa 1976) 2007;32:S91–S100. 6. Lenssinck ML, Frijlink AC, Berger MY, et al. Effect of bracing and other conservative interventions in the treatment of idiopathic scoliosis in adolescents: a systematic review of clinical trials. Phys Ther 2005;85:1329–1339. 7. Maruyama T. Bracing adolescent idiopathic scoliosis: a systematic review of the literature of effective conservative treatment looking for end results 5 years after weaning. Disabil Rehabil 2008;30:786–791. 8. el-Sayyad M, Conine TA. Effect of exercise, bracing and electrical surface stimulation on idiopathic scoliosis: a preliminary study. Int J Rehabil Res 1994;17:70–74. 9. Bunge EM, Habbema JD, de Koning HJ. A randomised controlled trial on the effectiveness of bracing patients with idiopathic scoliosis: failure to include patients and lessons to be learnt. Eur Spine J 2010;19:747–753. 10. Weinstein SL, Dolan LA, Wright JG, Dobbs MB. Effects of bracing in adolescents with idiopathic scoliosis. N Engl J Med 2013;369:1512–1521.
THE BONE & JOINT JOURNAL
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
11. Dolan LA, Sabesan V, Weinstein SL, Spratt KF. Preference assessment of recruitment into a randomized trial for adolescent idiopathic scoliosis. J Bone Joint Surg [Am] 2008;90-A:2594–2605. 12. Millat B, Borie F, Fingerhut A. Patient’s preference and randomization: new paradigm of evidence-based clinical research. World J Surg 2005;29:596–600. 13. Grufferman S. Complexity and the Hawthorne effect in community trials. Epidemiology 1999;10:209–210. 14. Pucci E, Belardinelli N, Signorino M, Angeleri F. Patients’ understanding of randomised controlled trials depends on their education. BMJ 1999;318:875. 15. Negrini S, Carabalona R. Social acceptability of treatments for adolescent idiopathic scoliosis: a cross-sectional study. Scoliosis 2006;1:14. 16. Fielding LP, Grace R, Hittinger R. Patients who are eligible but not randomised should be included as additional comparative arm in study. BMJ 1999;318:874–875. 17. Risser JC. The classic: the iliac apophysis: an invaluable sign in the management of scoliosis. Clin Orthop Relat Res 2010;468:643–653. 18. Keynan O, Fisher CG, Vaccaro A, et al. Radiographic measurement parameters in thoracolumbar fractures: a systematic review and consensus statement of the spine trauma study group. Spine 2006;31:E156–E165.
981
23. Li M, Wong MS, Luk KD, Wong KW, Cheung KM. Time-dependent response of scoliotic curvature to orthotic intervention: when should a radiograph be obtained after putting on or taking off a spinal orthosis? Spine (Phila Pa 1976) 2014;39:1408– 1416. 24. Asher M, Min Lai S, Burton D, Manna B. The reliability and concurrent validity of the scoliosis research society - patient questionnaire for idiopathic scoliosis. Spine 2003;28:63–69. 25. Spielberger CD. State-Trait Anxiety Inventory for adults: Sampler set: Manual, test booklet and scoring key. Redwood City, CA: Mind Garden, Inc; 1983. 26. Beck AT, Steer RA. G.K. B. BDI-II Manual. New York: The Psychological Corporation; 1996. 27. Cheung KM, Senkoylu A, Alanay A, et al. Reliability and concurrent validity of the adapted Chinese version of Scoliosis Research Society-22 (SRS-22) questionnaire. Spine (Phila Pa 1976) 2007;32:1141–1145. 28. Shek DT. The Chinese version of the State-Trait Anxiety Inventory: its relationship to different measures of psychological well-being. J Clin Psychol 1993;49:349–358. 29. Byrne BM, Stewart SM, Lee PWH. Validating the Beck Depression Inventory-II for Hong Kong Community Adolescents. Int J Test 2004;4:199–216. 30. Olschewski M, Schumacher M, Davis KB. Analysis of randomized and nonrandomized patients in clinical trials using the comprehensive cohort follow-up study design. Control Clin Trials 1992;13:226–239.
19. Richards BS, Bernstein RM, D’Amato CR, Thompson GH. Standardization of criteria for adolescent idiopathic scoliosis brace studies: SRS Committee on Bracing and Nonoperative Management. Spine (Phila Pa 1976) 2005;30:2068–2075.
31. Cheung KM, Cheng EY, Chan SC, Yeung KW, Luk KD. Outcome assessment of bracing in adolescent idiopathic scoliosis by the use of the SRS-22 questionnaire. Int Orthop 2007;31:507–511.
20. Lancaster GA, Dodd S, Williamson PR. Design and analysis of pilot studies: recommendations for good practice. J Eval Clin Pract 2004;10:307–312.
32. Reamy BV, Slakey B. Adolescent idiopathic scoliosis: review and current concepts. Am Fam Physician 2001;64:111–116.
21. Schiller JR, Thakur NA, Eberson CP. Brace management in adolescent idiopathic scoliosis. Clin Orthop Relat Res 2010;468:670–678.
33. Donzelli S, Zaina F, Negrini S. In defense of adolescents: they really do use braces for the hours prescribed, if good help is provided. Results from a prospective everyday clinic cohort using thermobrace. Scoliosis 2012;7:12.
22. No authors listed. Scoliosis and Kyphosis Brace Wear. OIC. http://ortho-institute.org/ (date last accessed 8 April 2015).
34. Lou E, Hill D, Hedden D, et al. An objective measurement of brace usage for the treatment of adolescent idiopathic scoliosis. Med Eng Phys 2011;33:290–294.
VOL. 97-B, No. 7, JULY 2015
An alternative to a randomised control design for assessing the efficacy and effectiveness of bracing in adolescent idiopathic scoliosis D. Y. T. Fong, K. M. C. Cheung, Y. W. Wong, W. Y. Cheung, I. C. Y. Fu, E.E. Kuong, K. C. Mak, M. To, D. Samartzis, K. D. K. Luk From The University of Hong Kong, Hong Kong, China D. Y. T. Fong, PhD, Associate Professor, School of Nursing I. C. Y. Fu, DN, Lecturer, School of Nursing The University of Hong Kong, 21 Sassoon Road, Pokfulam, Hong Kong SAR, China. K. M. C. Cheung, MBBS, MD, FRCS, FHKCOS, FHKAM (Orth), Clinical Professor and Head, Department of Orthopaedics & Traumatology Y. W. Wong, MBBS, FRCSE, FHKCOS, FHKAM (Orth), Consultant, Department of Orthopaedics & Traumatology W. Y. Cheung, MBBS, FRCSE, FHKAM, FHKCOS, Clinical Assistant Professor, Department of Orthopaedics & Traumatology E.E. Kuong, MBBS, Associate Consultant, Department of Orthopaedics & Traumatology K. C. Mak, MBBS, Associate Consultant, Department of Orthopaedics & Traumatology M. To, MBBS, FRCSEd, FHKCOS, FHKAM, Clinical Assistant Professor, Department of Orthopaedics & Traumatology D. Samartzis, DSc, Assistant Professor, Department of Orthopaedics & Traumatology K. D. K. Luk, MBBS, MChOrth, FRCSEd, FRCSGlas, FRACS, FHKAM, Chair Professor, Department of Orthopaedics & Traumatology The University of Hong Kong, 5/F, Professorial Block, Queen Mary Hospital, Pokfulam, Hong Kong SAR, China. Correspondence should be sent to Dr K. D. K. Luk; e-mail: [email protected] ©2015 The British Editorial Society of Bone & Joint Surgery doi:10.1302/0301-620X.97B7. 35147 $2.00 Bone Joint J 2015;97-B:973–81. Received 18 September 2014; Accepted after revision 5 March 2015
VOL. 97-B, No. 7, JULY 2015
Randomised controlled trials (RCTs) that assessed the efficacy of bracing for adolescent idiopathic scoliosis have suffered from small sample sizes, low compliance and lack of willingness to participate. The aim of this study was to assess the feasibility of a comprehensive cohort study for evaluating both the efficacy and the effectiveness of bracing in patients with adolescent idiopathic scoliosis. Patients with curves at greater risk of progression were invited to join a randomised controlled trial. Those who declined were given the option to remain in the study and to choose whether they wished to be braced or observed. Of 87 eligible patients (5 boys and 63 girls) identified over one year, 68 (78%) with mean age of 12.5 years (10 to 15) consented to participate, with a mean follow-up of 168 weeks (0 to 290). Of these, 19 (28%) accepted randomisation. Of those who declined randomisation, 18 (37%) chose a brace. Patients who were more satisfied with their image were more likely to choose bracing (Odds Ratio 4.1; 95% confidence interval 1.1 to 15.0; p = 0.035). This comprehensive cohort study design facilitates the assessment of both efficacy and effectiveness of bracing in patients with adolescent idiopathic scoliosis, which is not feasible in a conventional randomised controlled trial. Cite this article: Bone Joint J 2015; 97-B:973–81
Bracing has a long history in the treatment of adolescent idiopathic scoliosis (AIS), and it is recognised that patients who present radiographically with coronal curves are at high risk of progression.1 However, reports are mixed on whether bracing reduces curve progression and the need for surgery.2 A prospective cohort study by the Scoliosis Research Society (SRS) reported that bracing could prevent curve progression more effectively than observation alone.3 However, this study was not randomised with concurrent controls. In addition, there have been five systematic reviews or meta-analyses of bracing, including one underpowered randomised controlled trial (RCT).2,4-8 There have been only two large multicentre RCTs that have assessed the efficacy of bracing for AIS,9,10 but both encountered recruitment difficulties.9,11 Although an RCT design is regarded as the ideal for establishing efficacy, it has limitations, especially when used to assess the effectiveness of bracing in practice. This is because no-one can be blinded to bracing, and their preferences and expectations may influence the effectiveness of of this treatment.12,13 Cultural, social and economic factors may also affect decisions by both patients and parents to participate in an RCT.14 There have been reports
of poor participation in RCTs, which severely limit the generalisability of study findings.9 Therefore, it is important to find out what both patients and their families think of the available therapeutic options for AIS, which cannot be addressed in a traditional RCT design.15 We have proposed an alternative study design to evaluate both the efficacy and the effectiveness of bracing over observation. In order to do this we conducted a feasibility study to obtain preliminary information.
Patients and Methods We designed a comprehensive cohort study (CCS) in which treatment with bracing or observation was determined either by randomisation or by the preference of the patients and/ or the parents. Eligible patients were followedup every four months for a minimum of 12 months. Ethical approval was granted by a local institutional research board (NCT00989495). The CCS design is a new paradigm of evidencebased clinical research that combines an RCT and a preference-based controlled trial.12 The treatment allocation plan is shown in Figure 1. The trial examines the treatment efficacy, as in a traditional RCT (Group I vs Group II in Fig. 1); it also addresses the issue of effectiveness, by 973
974
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
Assessed for eligibility by reading patient records (n = 6479)
Assessed for eligibility by orthopaedic surgeons (n = 423) Not meeting inclusion criteria (n = 336) Eligibility confirmed (n = 87) Declined to participate (n = 19) Yes
Group I:
Group II:
(Brace) (n = 13) • Braced (n = 11) • Did not brace (n = 2; withdrawn)
(Observation) (n = 6) • Observed (n = 6)
• Followed (n = 11) • Withdrawn (n = 2); no reason)
Visit 2
Visit 3
Baseline
Visit 1
No
Consent to randomisation?
Group III: Brace (Brace) (n = 9) • Braced (n = 9)
Group IV: Observation (Observation) (n = 19) • Observed (n = 18) • Braced (n = 1); made in Nanjing by parents)
Group V: Brace (Observation) (n = 9) • Braced (n = 9)
Group VI: Observation (Brace) (n = 12) • Observed (n = 12)
• Followed (n = 6) • Withdrawn (n = 0)
• Followed (n = 9) • Withdrawn (n = 0)
• Followed (n = 18) • Withdrawn (n = 1); no reason)
• Followed (n = 9) • Withdrawn (n = 1); mother refused to complete questionnaires)
• Followed (n = 12) • Withdrawn (n = 0)
• Followed (n = 11) • Withdrawn (n = 2)
• Followed (n = 6) • Withdrawn (n = 0)
• Followed (n = 8) • Withdrawn (n = 1); no reason)
• Followed (n = 18) • Withdrawn (n = 1)
• Followed (n = 9) • Withdrawn (n = 2); father refused to complete questionnaires)
• Followed (n = 12) • Withdrawn (n = 0)
• Followed (n = 11) • Withdrawn (n = 2)
• Followed (n = 6) • Withdrawn (n = 0)
• Followed (n = 8) • Withdrawn (n = 1)
• Followed (n = 18) • Withdrawn (n = 1)
• Followed (n = 9) • Withdrawn (n = 2)
• Followed (n = 12) • Withdrawn (n = 0)
Fig. 1 Flow chart of patient disposition. For patients who consented to be randomised, the treatment in brackets indicates what was decided by the randomisation. For patients who declined randomisation and chose their treatments, the treatments in brackets indicate what would have been given if randomised, which may not necessarily be the same as the treatments they actually received.
assessing the impact of the treatment preference (Group I vs Group II and Groups III + V vs Groups IV + VI in Fig. 1). Because of the concurrent assessment of efficacy and effectiveness, the inclusion of non-randomised subjects has even been suggested in all RCTs to assess external validity.16 Patients aged ten years or older with Risser17 sign 0 to 2, and diagnosed with AIS with either a Cobb18 angle of 20° to < 25° and at least 5° deterioration over the past four months, or a Cobb angle of 25° to < 30°, were invited to participate. Exclusion criteria were patients with a history of treatment for AIS; who were more than year postmenarche; who had a physical or mental disability that interfered with adherence to bracing; who had a diagnosed musculoskeletal or developmental illness that might be responsible for the spinal curvature; who had a structural thoracic scoliosis with the apex above T7; who had difficulty in understanding or completing the study questionnaires; or who were unlikely to comply with the study protocol. These criteria conformed with those suggested by the SRS for bracing studies.19 We planned to recruit at least
ten patients into each of the two RCT groups and at least five into each of the other four groups (Fig. 1). Hence, a total of at least 40 patients were to be recruited for the pilot study.20 Interventions. For patients who underwent bracing, a custom-made Boston-type of thoracolumbar sacral orthosis (TLSO)21 was moulded under traction by an orthotist. The brace extended anteriorly from the pubis to the xiphoid process. Patients were asked to wear the brace for a minimum of 20 hours per day, and the brace straps were marked to ensure that correct pressure was applied. Patients under observation only returned for assessment as scheduled. Study procedures. Before patient recruitment, a treatment allocation sequence was generated by means of block randomisation, with block size randomised as four or eight. The allocation sequence was concealed using opaque envelopes, each containing the allocation code of one patient. Patients were recruited from the clinic of a children’s hospital largely devoted to orthopaedic patients. Before each clinic session, a recruitment team composed of a THE BONE & JOINT JOURNAL
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
975
Table I. The sociodemographic characteristics of the 68 study patients and their parents
Mean age (yrs) (standard deviation SD) Gender Male (n, %) Female (n, %) Educational level Grade 5 to 6 (n, %) Grade 7 to 11 (n, %) Mean height (cm) (SD) Mean weight* (kg) (SD) Mean BMI* (kg/m2) (SD) Mean arm span (cm) (SD) Mean monthly household income (HK$) < 2000 (n, %) 2000 to 5999 (n, %) 600 to 9999 (n, %) 10 000 to 14 999 (n, %) 15 000 to 19 999 (n, %) 20 000 to 24 999 (n, %) 25 000 to 29 999 (n, %) 30 000 to 39 999 (n, %) 40 000 to 59 999 (n, %) ≥ 60 000 (n, %) Mean gestational age (wks) (SD) Mean maternal age (yrs) (SD) Breastfeeding No (n, %) Yes (n, %) Do not know/Missing (n, %) Smoking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Alcohol drinking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Family history of scoliosis No (n, %) Yes (n, %) Do not know/missing (n, %)
RCT component (n = 19)
Preference component (n = 49)
p-value
12.4 (1.1)
12.6 (1.1)
0.383
1 (5.3) 18 (94.7)
4 (8.2) 45 (91.8)
0.684
7 (36.8) 12 (63.2) 150.9 (8.0) 38.3 (7.9) 16.7 (2.6) 152.2 (10.2)
18 (36.7) 31 (63.3) 154.8 (5.6) 41.9 (8.6) 17.4 (3.1) 155.5 (7.0)
2 (10.5) 1 (5.3) 2 (10.5) 3 (15.8) 4 (21.1) 3 (15.8) 0 (0) 1 (5.3) 1 (5.3) 2 (10.5) 39.7 (1.4) 30.3 (4.3)
1 (2.0) 2 (4.1) 5 (10.2) 3 (6.1) 11 (22.4) 6 (12.2) 3 (6.1) 5 (10.2) 5 (10.2) 8 (16.3) 39.4 (1.2) 30.0 (5.0)
13 (68.4) 6 (31.6) 0 (0)
34 (69.4) 13 (26.5) 2 (4.1)
13 (68.4) 6 (31.6) 0 (0)
27 (55.1) 20 (40.8) 2 (4.1)
12 (63.2) 7 (36.8) 0 (0)
26 (59.2) 20 (40.8) 3 (6.1)
14 (73.7) 1 (5.3) 4 (21.1)
29 (59.2) 8 (16.3) 12 (24.5)
0.993
0.039 0.125 0.404 0.138 0.096
0.393 0.805 0.750
0.411
0.622
0.223
* Three missing standard deviation; BMI, body mass index
SD,
research assistant and two of the authors (DF, IF) screened all the clinical records and identified potential patients. These patients were seen by orthopaedic surgeons to assess their eligibility. Patients who were suitable were referred to the recruitment team, who explained the study details and sought informed consent. Each eligible patient was invited to participate in an RCT. Those who declined were given an option to stay in the study and choose whether they wished to be braced or observed. The consent form was signed together with one of their accompanying guardians or parents. The 60 patients with the best study compliance received about $25 (USD) to offset their travel expenses. Braced patients were telephoned at least one day before their next clinical visit to remind them to remove the brace at least 24 hours before their visit. In the literature, the length of time out-of-brace varies, but at least 24 hours free VOL. 97-B, No. 7, JULY 2015
from the brace would ensure full relaxation of the spine for a more realistic measurement of the true curvature.22,23 Measurements. The sociodemographic details of all patients were recorded. At all clinical visits, height (cm), weight (kg), arm span (cm), menarche status and exercise habits were measured. All participants self-completed the Scoliosis Research Society-22 (SRS-22) questionnaire, the State-Trait Anxiety Inventory (STAI) and the Beck Depression Inventory version II (BDI-II).24-26 These tools have been shown to have good validity and reliability in Chinese subjects.27-29 In addition, standing posteroanterior out-ofbrace radiographs of the full spine were taken. An orthopaedic surgeon (EK) performed blinded assessment of the radiographs. Only one observer was used for assessment of the radiographs to avoid interobserver differences, as recommend by the SRS.19 Braced patients were also given a
976
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
Table II. The curvature and quality of life measures of the patients and their parents RCT Component (n = 19)
Cobb angle (º) State anxiety Patient Mother Father Trait anxiety Patient Mother Father Beck Depression Inventory Patient Mother Father SRS-22 Function Pain Mage Mental health Management Subtotal Total
Preference component (n = 49)
n
Mean (SD)
n
Mean (SD)
p-value
18
25.4 (3.5)
48
25.5 (3.5)
0.938
19 16 4
46.0 (11.2) 48.9 (8.2) 45.3 (7.2)
49 38 15
43.9 (8.9) 49.8 (12.8) 44.7 (8.3)
0.404 0.790 0.905
19 16 4
42.2 (7.8) 45.4 (5.8) 42.3 (5.4)
49 38 15
41.7 (8.5) 45.0 (7.5) 39.4 (7.4)
0.826 0.845 0.462
19 16 4
5.6 (6.6) 4.1 (4.2) 6.3 (1.0)
48 36 15
4.5 (4.6) 6.6 (6.5) 3.3 (5.1)
0.420 0.173 0.269
18 18 18 18 15 18 15
4.9 (0.1) 4.7 (0.4) 3.8 (0.6) 3.8 (0.6) 3.8 (0.6) 4.4 (0.3) 4.3 (0.3)
43 43 43 43 34 43 34
4.9 (0.2) 4.7 (0.4) 3.9 (0.6) 4.1 (0.5) 3.5 (0.6) 4.4 (0.3) 4.3 (0.3)
0.635 0.968 0.781 0.435 0.130 0.711 0.707
RCT, Randomised controlled trial; standard deviation; SRS, Scoliosis Research Society
SD,
diary to record the daily number of hours in the brace and asked to return the completed diaries at each follow-up visit. An accompanying parent of each patient also selfcompleted the STAI and BDI-II. Statistical analysis. Quality of life measures were scored in accordance with the corresponding scoring manuals.25,26 Specifically, the SRS-22 has domains scored as means in the range of 1 to 5. If fewer than half of the items in a domain had responses, the skipped items were scored by the mean of the other items in the domain. Otherwise, the domain score was recorded as missing. Determinants of acceptance to randomisation were examined by comparing groups I + II and III + IV + V + VI (described in Fig. 1) by a stepwise logistic regression on all baseline measurements. The effects of bracing on quality of life and SRS-22 were assessed by a linear mixed effects model to account for repeated measurements over time, whereas those at risk of ≥ 6º progression were assessed by Cox’s regression analysis. All significance tests were two-tailed, and a p-value < 0.05 was deemed significant. Analysis was performed using SAS v 9.2 software (SAS Institute, Cary, North Carolina).
Results Between 9 March 2009 and 8 March 2010, a total of 87 eligible AIS patients were approached and 68 (78%) consented to participate in the study. About six patients per month were recruited. Among the 19 refusals, 16 (84%) were made by the parents (11 mothers and five fathers) and three were made by the children (Fig. 1). Parents refused to participate because they were unwilling to be involved in a
research study (n = 7); they wanted to decide at the next visit (n = 6); they were in a hurry (n = 2); or they thought a retrospective study was possible without consent (n = 1). One newly diagnosed girl was too emotionally unstable to participate. Another girl whose mother consented to randomisation was afraid of bracing and refused to participate. The other child (a boy) refused participation because he was unwilling to complete the questionnaires. Of the 68 study patients (five boys, 63 girls), 19 (28%) consented to be randomised (Tables I and II). The participation rate in the RCT was 19/87 (22%). Among the 49 patients who chose their own treatment, 18 (37%) chose bracing (Tables III and IV). Taller patients (Odds Ratio (OR) 0.90, 95% CI 0.81 to 0.995; p = 0.040) or those who had a higher monthly household income (OR 0.75, 95% CI 0.57 to 0.98; p = 0.032) preferred to choose their treatment (Table V). Among patients who had their treatment preferences, those with one unit score higher in the domain ‘satisfaction with self-image’ were three-times more likely to choose bracing (OR 4.1, 95% CI 1.1 to 15.0; p = 0.035) (Table V). Patients were followed for a maximum of one year on the self-reported assessments, but the follow-up of their curvature measurements was extended to January 2015. The mean duration of follow-up was 168 weeks (0 to 290), and six patients withdrew from the study within the first year of follow-up. The parents of two patients refused to continue with the study because of unwillingness to complete the questionnaires. Another three did not continue follow-up at the specialist hospital after recruitment, and never reattended. The other two patients did not provide reasons THE BONE & JOINT JOURNAL
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
977
Table III. The sociodemographic characteristics of patients who desired a certain treatment
Mean age (yrs) (SD) Gender Male (n, %) Female (n, %) Educational level Grade 5 to 6 (n, %) Grade 7 to 11 (n, %) Mean height (cm) (SD) Mean weight* (kg) (SD) Mean BMI* (kg/m2) (SD) Mean arm span (cm) (SD) Monthly household income (HK$) < 2000 (n, %) 2000 to 5999 (n, %) 6000 to 9999 (n, %) 10 000 to 14 999 (n, %) 15 000 to 19 999 (n, %) 20 000 to 24 999 (n, %) 25 000 to 29 999 (n, %) 30 000 to 39 999 (n, %) 40 000 to 59 999 (n, %) ≥ 60 000 (n, %) Mean gestational age (wks) (SD) Mean maternal age (yrs) (SD) Breastfeeding No (n, %) Yes (n, %) Do not know/missing (n, %) Smoking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Alcohol drinking among family members No (n, %) Yes (n, %) Do not know/missing (n, %) Family history of scoliosis No (n, %) Yes (n, %) Do not know/missing (n, %)
Brace (n = 18)
Observation (n = 31)
p-value
12.4 (0.9)
12.7 (1.2)
0.275 0.616
1 (5.6) 17 (94.4)
3 (9.7) 28 (90.3)
7 (38.9) 11 (61.1) 153.7 (5.4) 41.0 (8.7) 17.3 (2.9) 154.6 (7.1)
11 (35.5) 20 (64.5) 155.4 (5.8) 42.5 (8.6) 17.5 (3.3) 156.1 (7.1)
1 (5.6) 0 (0) 3 (16.7) 1 (5.6) 1 (5.6) 3 (16.7) 2 (11.1) 3 (16.7) 2 (11.1) 2 (11.1) 39.7 (1.3) 28.2 (5.1)
0 (0) 2 (6.5) 2 (6.5) 2 (6.5) 10 (32.3) 3 (9.7) 1 (3.2) 2 (6.5) 3 (9.7) 6 (19.4) 39.3 (1.1) 31.1 (4.7)
13 (72.2) 5 (27.8) 0 (0)
21 (67.7) 8 (25.8) 2 (6.5)
8 (44.4) 10 (55.6) 0 (0)
19 (61.3) 10 (32.3) 2 (6.5)
9 (50.0) 8 (44.4) 1 (5.6)
17 (54.8) 12 (38.7) 2 (6.5)
11 (61.1) 3 (16.7) 4 (22.2)
18 (58.1) 5 (16.1) 8 (25.8)
0.812
0.311 0.567 0.802 0.478 0.953
0.267 0.056 0.989
0.159
0.708
0.982
* Two missing SD, standard deviation; BMI, body mass index; HK$, Hong Kong dollars
for their withdrawal. Of 29 braced patients, 27 (93%) returned their diaries. The overall mean number of hours in the brace during the first year was 17.0 per day (95% CI 15.5 to 18.6), which was not significantly different between those randomised and those who chose to be braced (p = 0.634, by t-test). In the RCT arm there were no significant differences in anxiety and depression scores between braced and the observed patients (p > 0.212, by linear mixed effects model) (Fig. 2). In the preference arm no significant differences were found (p > 0.202, by linear mixed effects model), except that mothers of the braced patients had their depression level reduced significantly more than those of the observed patients (p = 0.031, by linear mixed effects model). Figure 3 similarly assesses the effects of bracing on the SRS-22. In both the RCT and the preference arms VOL. 97-B, No. 7, JULY 2015
braced patients scored significantly lower in the function, image, subtotal and total scales (p ≤ 0.042, by linear mixed effects model). In addition, braced patients in the preference arm also had significantly lower mental health scores (p = 0.030, by linear mixed effects model). A total of 33 patients had ≥ 6º of curve progression during follow-up, eight of whom (five braced and three observed) had treatment by randomisation and 25 (eight braced and 17 observed) chose their treatment. In the RCT arm the relative risk (RR) of curve progression of ≥ 6 º with bracing was 1.43 (95% CI 0.27 to 7.61, p = 0.679). In the preference arm the corresponding RR was 0.27 (95% CI 0.10 to 0.75, p = 0.012). There was a trend towards a more profound effect of bracing in the preference arm than in the RCT arm (ratio of RRs = 0.20, 95% CI 0.03 to 1.33; p = 0.095).
978
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
Table IV. The curvature and quality of life measures of patients who desired a certain treatment Brace (n = 18)
Cobb angle (º) State anxiety Patient Mother Father Trait anxiety Patient Mother Father Beck depression inventory Patient Mother Father SRS-22 Function Pain Image Mental health Management Subtotal Total SD,
Observation (n = 31)
n
Mean (SD)
n
Mean (SD)
p-value
18
26.4 (4.1)
30
24.9 (3.0)
0.160
18 15 4
44.9 (10.2) 31 53.5 (13.1) 23 37.5 (12.5) 11
43.3 (8.1) 47.5 (12.2) 47.4 (4.7)
0.532 0.161 0.141
18 15 4
41.6 (9.1) 31 45.33 (8.4) 23 37.5 (9.3) 11
41.7 (8.2) 44.7 (7.0) 40.1 (7.0)
0.968 0.808 0.539
17 15 4
4.8 (S4.8) 9.0 (8.3) 2.5 (3.8)
31 21 11
4.3 (4.5) 4.9 (4.3) 3.6 (5.6)
0.716 0.079 0.717
15 15 15 15 10 15 10
5.0 (0.1) 4.7 (0.3) 4.1 (0.5) 4.2 (0.6) 3.8 (0.7) 4.5 (0.3) 4.4 (0.3)
28 28 28 28 24 28 24
4.9 (0.2) 4.7 (0.5) 3.7 (0.6) 4.1 (0.5) 3.5 (0.6) 4.4 (0.3) 4.3 (0.3)
0.259 0.712 0.035 0.400 0.213 0.112 0.232
standard deviation; SRS, Scoliosis Research Society
Table V. Determinants of acceptance of randomisation and choosing brace treatment, by stepwise logistic regression
Determinants of acceptance of randomisation (n = 68) Patient’s height (cm) Monthly household income (HK$) Determinants of choosing brace treatment (n = 49) Satisfaction with self-image
Odds ratio
95% CI
p-value
0.90 0.75
(0.81 to 0.995) (0.57 to 0.98)
0.040 0.032
4.1
(1.10 to 15.0)
0.035
CI, confidence interval
RCT arm absolute p-values: STAI_state (patient), p = 0.645; STAI_trait (patient), p = 0.212; BDI (patient), p = 0.583; STAI_state (mother), p = 0.719; STAI_trait (mother), p = 0.272; BDI (mother), p = 0.645. Preference arm absolute p-values: STAI_state (patient), p = 0.293; STAI_trait (patient), p = 0.202; BDI (patient), p = 0.229; STAI_state (mother), p = 0.980; STAI_trait (mother), p = 0.576; BDI (mother), p = 0.031. RCT arm absolute p-values: Function, p = 0.042; Image, p = 0.035; Subtotal, p = 0.011; Total, p = 0.018. Preference arm absolute p-values: Function, p = 0.001; Image, p < .001; Subtotal, p < 0.001; Total, p = 0.002.
Discussion The CCS design addresses the concerns raised by traditional RCTs for bracing, owing to the low acceptance of treatment decided by randomisation. The participation rate in our RCT component was 21%, the same as that in a Dutch RCT but lower than the 33% reported in a hypothetical setting.9,11 Refusal to consent to enter an RCT as a result of patients’ preference as to their treatment was
observed in this study. The effect of bracing on the risk of ≥ 6º progression in the preference arm was larger than that in the RCT arm, although the difference did not reach statistical significance owing to the small sample size. The potential influence of such preference on brace effects can be addressed by including the preference component under our CCS, which is what a traditional RCT cannot accomplish. The analysis of a CCS should begin with an assessment of efficacy of bracing by comparing the brace (Group I) and the observation (Group II) groups under the RCT component. Then the demographic and socioeconomic status of patients between the RCT (Groups I + II) and the preference (Groups III + IV + V + VI) components are compared with identify determinants of treatment preference. Finally, the impact of treatment preference on bracing effects is examined by assessing the preference by treatment interaction, i.e. (Groups I vs Group II) and (Groups III + V vs Groups IV + VI).30 When there is no indication of an influence of treatment preference, a more reliable estimate of bracing effects based on patients in the two components can be obtained. THE BONE & JOINT JOURNAL
50 40
60 50 40 30
20
20
BDI (patients)
50 40 30 20
SRS_Function
4
4
3 Observation Brace
Observation Brace
50 40 30
4 3
20
Observation Brace
2
60 50 40
70 60 50 40
70 60 50 40 30
1 5
4
4
3 Observation Brace
Observation Brace
70 60 50 40
20 60
Observation Brace
Observation Brace
40 30 20
SRS_subtotal -
50 BDI (mother)
50
40 30 20
4 3 Observation Brace
2
2
3
0
Visit
1
2
3
Visit
Fig. 2 Graphs showing the longitudinal change in mean anxiety and depression scores by treatment group.
SRS_Total
1
5
4 3 Observation Brace
VOL. 97-B, No. 7, JULY 2015
Observation Brace
2 1
4 3 Observation Brace
2
1 5
1 5
4
4
3 Observation Brace
2
The treatment decision was mostly made by the patients’ parents, and most chose observation. Parents of taller patients think their children have little potential to grow taller and so they were more determined to choose observation. On the other hand, parents with greater financial resources were often more educated and had a better understanding of the risks and benefits of bracing. They were
3
5
SRS_Total
0 0
Observation Brace
4
1
10
0
3
1 5
2 10
Observation Brace
2
1 5
20 80
30
20 60
3
1 5
2
SRS_Management
STAI_Trait (mothers)
Observation Brace
4
Observation Brace
30
20 80
Observation Brace
2
SRS_MH
70
0 80
SRS_MH
Observation Brace
3
1 5
10
0 80
Observation Brace
2
1 5
60
Observation Brace
1 5
2
30
3
1 5
SRS_Pain
60
Observation Brace
70
4
2
Observation Brace
SRS_Image
Observation Brace
3
SRS_Management
20 80
70
4
2
20 80
STAI_State (mothers)
STAI_State (mothers)
40 30
30
STAI_Trait (mothers)
50
30
10
BDI (mother)
60
979
5
SRS_subtotal -
40
SRS_Function
50
5
Observation Brace
70
SRS_Pain
60
80
SRS_Image
Observation Brace
70
60 BDI (patients)
STAI_State (patients)
80
STAI_Trait (patients)
STAI_Trait (patients)
STAI_State (patients)
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
3 Observation Brace
2
1
1 0
1
2
3
0
Visit
1
2
3
Visit
Fig. 3 Graphs showing the longitudinal change in mean SRS-22 component scores by treatment group.
more prone to express preferences and to refuse allocation of treatment by randomisation.
980
D. Y. T. FONG, K. M. C. CHEUNG, Y. W. WONG, W. Y. CHEUNG, I. C. Y. FU, E.E. KUONG, K. C. MAK, M. TO, D. SAMARTZIS, K. D. K. LUK
It is interesting to note that greater patient satisfaction with their image could influence their parents to choose bracing. A main consideration of parents was whether their children would accept bracing, which has previously been shown to reduce the satisfaction of patients with their image.31 It could be that patients who had greater satisfaction with their image have greater tolerance of their appearance and thus are more open to bracing. On the other hand, the satisfaction level with image in patients who chose observation, albeit lower than that in patients who chose bracing, was still good. These patients did not exhibit significant dissatisfaction with their curvature, to the extent that it would make them choose to be braced to improve their appearance. Several important aspects must be considered when planning for the full-scale CCS. First, during patient recruitment, it is important that all orthopaedic surgeons and patient recruiters remain impartial when offering advice to patients. For example, it has been the preference of the surgeons in our clinic to brace skeletally immature patients with curves ≥ 30° because of the perceived higher risk of progression.32 These surgeons have found it difficult to stay impartial when offering advice to the group in this study. Hence it was decided to restrict the inclusion criterion to curves that < 30°, so that the recruiters could comfortably explain that there is no good evidence available to support either bracing or observation. Second, there were unbalanced numbers of patients across the different treatment groups. Balanced group sizes would result in smaller standard errors and hence give larger power to detect an effect. Therefore, block randomisation stratified by whether patients consent to randomisation or not is recommended. Third, it is preferable to allow full relaxation of the spine to facilitate a realistic measure of the curvature. Fourth, temperature sensors could give a more accurate record of the actual hours of brace wear,33 However, there is still no guarantee that the strap tensions were appropriately tightened. In fact, pressure sensors have been developed to monitor both the duration and the quality of bracing.34 Fifth, we did not find that the financial incentive added to the participation of parents in our study, as parents were predominantly more concerned about which was the best treatment for their children, rather than any monetary compensation. Sixth, several parents were unwilling to complete the questionnaires covering the psychosocial domains. Strategies would be formulated to enhance participants’ compliance. A study by Weinstein et al10 with a similar design, which began recruiting subjects after our initial study recruitment, was terminated prematurely owing to the effect of bracing over observation in reducing curve progression. As in our study, Weinstein et al had difficulty in recruiting patients for the RCT arm of their study, and they noted the inclination of patients to choose the treatment type option. However, their study population was based on a heterogeneous
Western population and not an Asian one, which may also present with specific ethnic and cultural considerations related to study and treatment selection, such as degree of cosmetic concern and risk of skin irritation, and hence affect the outcomes. Our study, in which it was impossible to blind patients, parents and physicians to brace treatment, has revealed that parents have a strong preference to choose their desired treatment, which limits the role of traditional RCTs. We have successfully completed a feasibility study of a CCS design that allows the assessment of the impact of preference on the effectiveness of bracing. A full-scale trial is now being planned, which can be further used to address the efficacy and effectiveness between different populations. Author contributions: D. Y. T. Fong: Research design, Case recruitment, Data collection, Analysis, Writing paper. K. M. C. Cheung: Research design, Case recruitment, Data collection, Analysis, Writing the paper. Y. W. Wong: Research design, Case recruitment, Data collection, Analysis, Writing the paper. W. Y. Cheung: Case recruitment, Data collection, Writing the paper. I. C. Y. Fu: Case recruitment, Data collection, Writing the paper. E. E Kuong: Case recruitment, Data collection, Writing the paper. K. C. Mak: Case recruitment, Data collection, Writing the paper. M. To: Case recruitment, Data collection, Writing the paper. D. Samartzis: Analysis, Writing the paper. K. D. K. Luk: Research design, Case recruitment, Data collection, Analysis, Writing the paper. The work was financially supported by the Scoliosis Research Society. We thank Sister Leung, Ms J. Lam, Ms SY Lam and all nursing staff at the Duchess of Kent Children’s Hospital for facilitating the subject recruitment. We also thank Ms NL Kwok for recruiting the patients and coordinating the study, as well as Ms P. Wan and Mr G. Chan for following up the patients. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This article was primary edited by S. P. F. Hughes and first proof edited by G. Scott.
References 1. Asher MA, Burton DC. Adolescent idiopathic scoliosis: natural history and long term treatment effects. Scoliosis 2006;1:2. 2. Maruyama T, Grivas TB, Kaspiris A. Effectiveness and outcomes of brace treatment: a systematic review. Physiother Theory Pract 2011;27:26–42. 3. Nachemson AL, Peterson LE. Effectiveness of treatment with a brace in girls who have adolescent idiopathic scoliosis. A prospective, controlled study based on data from the Brace Study of the Scoliosis Research Society. J Bone Joint Surg [Am] 1995;77-A:815–822. 4. Rowe DE, Bernstein SM, Riddick MF, et al. A meta-analysis of the efficacy of non-operative treatments for idiopathic scoliosis. J Bone Joint Surg [Am] 1997;79A:664–674. 5. Dolan LA, Weinstein SL. Surgical rates after observation and bracing for adolescent idiopathic scoliosis: an evidence-based review. Spine (Phila Pa 1976) 2007;32:S91–S100. 6. Lenssinck ML, Frijlink AC, Berger MY, et al. Effect of bracing and other conservative interventions in the treatment of idiopathic scoliosis in adolescents: a systematic review of clinical trials. Phys Ther 2005;85:1329–1339. 7. Maruyama T. Bracing adolescent idiopathic scoliosis: a systematic review of the literature of effective conservative treatment looking for end results 5 years after weaning. Disabil Rehabil 2008;30:786–791. 8. el-Sayyad M, Conine TA. Effect of exercise, bracing and electrical surface stimulation on idiopathic scoliosis: a preliminary study. Int J Rehabil Res 1994;17:70–74. 9. Bunge EM, Habbema JD, de Koning HJ. A randomised controlled trial on the effectiveness of bracing patients with idiopathic scoliosis: failure to include patients and lessons to be learnt. Eur Spine J 2010;19:747–753. 10. Weinstein SL, Dolan LA, Wright JG, Dobbs MB. Effects of bracing in adolescents with idiopathic scoliosis. N Engl J Med 2013;369:1512–1521.
THE BONE & JOINT JOURNAL
AN ALTERNATIVE TO A RANDOMISED CONTROL DESIGN FOR ASSESSING THE EFFICACY AND EFFECTIVENESS OF BRACING
11. Dolan LA, Sabesan V, Weinstein SL, Spratt KF. Preference assessment of recruitment into a randomized trial for adolescent idiopathic scoliosis. J Bone Joint Surg [Am] 2008;90-A:2594–2605. 12. Millat B, Borie F, Fingerhut A. Patient’s preference and randomization: new paradigm of evidence-based clinical research. World J Surg 2005;29:596–600. 13. Grufferman S. Complexity and the Hawthorne effect in community trials. Epidemiology 1999;10:209–210. 14. Pucci E, Belardinelli N, Signorino M, Angeleri F. Patients’ understanding of randomised controlled trials depends on their education. BMJ 1999;318:875. 15. Negrini S, Carabalona R. Social acceptability of treatments for adolescent idiopathic scoliosis: a cross-sectional study. Scoliosis 2006;1:14. 16. Fielding LP, Grace R, Hittinger R. Patients who are eligible but not randomised should be included as additional comparative arm in study. BMJ 1999;318:874–875. 17. Risser JC. The classic: the iliac apophysis: an invaluable sign in the management of scoliosis. Clin Orthop Relat Res 2010;468:643–653. 18. Keynan O, Fisher CG, Vaccaro A, et al. Radiographic measurement parameters in thoracolumbar fractures: a systematic review and consensus statement of the spine trauma study group. Spine 2006;31:E156–E165.
981
23. Li M, Wong MS, Luk KD, Wong KW, Cheung KM. Time-dependent response of scoliotic curvature to orthotic intervention: when should a radiograph be obtained after putting on or taking off a spinal orthosis? Spine (Phila Pa 1976) 2014;39:1408– 1416. 24. Asher M, Min Lai S, Burton D, Manna B. The reliability and concurrent validity of the scoliosis research society - patient questionnaire for idiopathic scoliosis. Spine 2003;28:63–69. 25. Spielberger CD. State-Trait Anxiety Inventory for adults: Sampler set: Manual, test booklet and scoring key. Redwood City, CA: Mind Garden, Inc; 1983. 26. Beck AT, Steer RA. G.K. B. BDI-II Manual. New York: The Psychological Corporation; 1996. 27. Cheung KM, Senkoylu A, Alanay A, et al. Reliability and concurrent validity of the adapted Chinese version of Scoliosis Research Society-22 (SRS-22) questionnaire. Spine (Phila Pa 1976) 2007;32:1141–1145. 28. Shek DT. The Chinese version of the State-Trait Anxiety Inventory: its relationship to different measures of psychological well-being. J Clin Psychol 1993;49:349–358. 29. Byrne BM, Stewart SM, Lee PWH. Validating the Beck Depression Inventory-II for Hong Kong Community Adolescents. Int J Test 2004;4:199–216. 30. Olschewski M, Schumacher M, Davis KB. Analysis of randomized and nonrandomized patients in clinical trials using the comprehensive cohort follow-up study design. Control Clin Trials 1992;13:226–239.
19. Richards BS, Bernstein RM, D’Amato CR, Thompson GH. Standardization of criteria for adolescent idiopathic scoliosis brace studies: SRS Committee on Bracing and Nonoperative Management. Spine (Phila Pa 1976) 2005;30:2068–2075.
31. Cheung KM, Cheng EY, Chan SC, Yeung KW, Luk KD. Outcome assessment of bracing in adolescent idiopathic scoliosis by the use of the SRS-22 questionnaire. Int Orthop 2007;31:507–511.
20. Lancaster GA, Dodd S, Williamson PR. Design and analysis of pilot studies: recommendations for good practice. J Eval Clin Pract 2004;10:307–312.
32. Reamy BV, Slakey B. Adolescent idiopathic scoliosis: review and current concepts. Am Fam Physician 2001;64:111–116.
21. Schiller JR, Thakur NA, Eberson CP. Brace management in adolescent idiopathic scoliosis. Clin Orthop Relat Res 2010;468:670–678.
33. Donzelli S, Zaina F, Negrini S. In defense of adolescents: they really do use braces for the hours prescribed, if good help is provided. Results from a prospective everyday clinic cohort using thermobrace. Scoliosis 2012;7:12.
22. No authors listed. Scoliosis and Kyphosis Brace Wear. OIC. http://ortho-institute.org/ (date last accessed 8 April 2015).
34. Lou E, Hill D, Hedden D, et al. An objective measurement of brace usage for the treatment of adolescent idiopathic scoliosis. Med Eng Phys 2011;33:290–294.
VOL. 97-B, No. 7, JULY 2015
