SPINE Volume 27, Number 18, pp 2036–2040
©2002, Lippincott Williams & Wilkins, Inc.
A Multicenter Study Analyzing the Relationship of a
Standardized Radiographic Scoring System of
Adolescent Idiopathic Scoliosis and the Scoliosis
Research Society Outcomes Instrument
Philip L. Wilson, MD,* Peter O. Newton, MD,* Dennis R. Wenger, MD,*
Thomas Haher, MD,† Andrew Merola, MD,† Larry Lenke, MD,‡ Thomas Lowe, MD,§
David Clements, MD,储 and Randy Betz, MD¶
Study Design. A multicenter study examining the as-
sociation between radiographic and outcomes measures
in adolescent idiopathic scoliosis.
Objectives. To evaluate the association between an
objective radiographic scoring system and patient quality
of life measures as determined by the Scoliosis Research
Society outcomes instrument.
Summary of Background Data. Although surgical cor-
rection of scoliosis has been reported to be positively
correlated with patient outcomes, studies to date have
been unable to demonstrate an association between ra-
diographic measures of deformity and outcomes mea-
sures in patients with adolescent idiopathic scoliosis.
Methods. A standardized radiographic deformity scor-
ing system and the Scoliosis Research Society outcome
tool were used prospectively in seven scoliosis centers to
collect data on patients with adolescent idiopathic scoli-
osis. A total of 354 data points for 265 patients consisting
of those with nonoperative or preoperative curves ⱖ10°,
as well as those with surgically treated curves, were an-
alyzed. Correlation analysis was performed to identify
significant relationships between any of the radiographic
measures, the Harms Study Group radiographic defor-
mity scores (total, sagittal, coronal), and the seven Scoli-
osis Research Society outcome domains (Total Pain, Gen-
eral Self-Image, General Function, Activity, Postoperative
Self-Image, Postoperative Function, and Satisfaction) as
well as Scoliosis Research Society outcomes instrument
total scores. Radiographic measures that were identified
as significantly correlated with Scoliosis Research Society
outcome scores were then entered into a stepwise regres-
sion analysis.
Results. The coronal measures of thoracic curve and
lumbar curve magnitude were found to be significantly
correlated with the Total Pain, General Self-Image, and
total Scoliosis Research Society scores (P ⬍ 0.0001). The
From the *Children’s Hospital and Health Center, San Diego, Califor-
nia, †St. Vincent’s Hospital, New York, New York, ‡Washington Uni-
versity School of Medicine, St. Louis, Missouri, §Woodridge Orthope-
dics, Wheat Ridge, Colorado, 储Temple University Hospital, and
¶Shriners Hospital for Children, Philadelphia, Pennsylvania.
Acknowledgment date: October 2, 2001.
First revision date: January 31, 2002.
Acceptance date: March 11, 2002.
The device(s)/drug(s) is/are FDA-approved or approved by correspond-
ing national agency for this indication.
Corporate/industry funds were received to support this work. One or
more of the author(s) has/have received or will receive benefits for
personal or professional use from a commercial party related directly
or indirectly to the subject of this manuscript, e.g., honoraria, gifts,
consultancies.
DOI: 10.1097/01.BRS.0000024169.56395.B4
2036
thoracic and upper thoracic curve magnitudes were also
correlated with General Function (P ⬍ 0.002). The “coro-
nal” subscore as well as the “total” score of the Harms
Study Group radiographic scoring system were also sig-
nificantly correlated with these Scoliosis Research Soci-
ety domain and total scores. No radiographic measures
taken after surgery were significantly correlated with the
postoperative domains of the Scoliosis Research Society
outcomes instrument. Stepwise regression analysis of
these radiographic measures as predictors of Scoliosis
Research Society scores resulted in adjusted R2 values of
0.03– 0.07 (P ⬍ 0.0001). Although these results show that
a significant association exists between the radiographic
Cobb angle measure of the scoliosis and the Scoliosis
Research Society outcomes scores, the low R2 values
indicate that variables other than the radiographic ap-
pearance of the deformity (e.g., psychosocial, functional)
must also be affecting these scores.
Conclusion. The Cobb angle measure of the major
deformity has a small, but statistically significant, corre-
lation with the reported Total Pain, General Self-Image,
and General Function as measured by the Scoliosis
Research Society outcomes instrument. None of the ra-
diographic measures in this population correlated with
postoperative domain scores of the Scoliosis Research
Society outcomes tool. [Key words: adolescent, outcome
assessment, scoliosis] Spine 2002;27:2036 –2040
The surgical treatment of adolescent idiopathic scoliosis
(AIS) is predicated on the basis of improving the truncal
deformity, limiting curve progression, and ultimately
leading to a greater quality of life. Although there is
potential for pulmonary compromise when curve mag-
nitudes reach 80 –100°, the indications for surgical inter-
vention in the majority of idiopathic scoliosis cases are
these aforementioned concerns.2,22 The decision to per-
form surgery for correction of idiopathic scoliosis is
largely based on objective radiographic measures. Tradi-
tionally, Cobb angle measures of the major coronal-
plane curves have been the standard means for quantify-
ing scoliotic deformity.13–15 Recently, in an effort to
more thoroughly quantify and address these spinal de-
formities, additional radiographic measures have been
systematically used.1 Despite the use of these radio-
graphic measures for pretreatment and posttreatment as-
sessment of AIS, there are few data correlating the objec-
tive radiographic measures to the stated treatment goals
 Scoliosis Research Society Outcomes Instrument • Wilson et al 2037

Table 1. Radiographic Measures instrument for patients evaluated or undergoing treatment for
AIS. This validated, disease-specific questionnaire consists of
Measure Points seven domains: Pain, General Self-Image, General Function,
Overall Level of Activity, Postoperative Self-Image, Postoper-
Coronal measurements
Thoracic curve 20
ative Function, and Satisfaction. The questionnaire is format-
Upper thoracic curve 5 ted in two sections. Section 1 consists of the first four domains
Lumbar curve 20 listed above and applies to pretreatment as well as posttreat-
Coronal C7 to CSL 5 ment patients. Section 2 applies to posttreatment patients only
Apical translation (apex to CSL) 4 and consists of the last three domains listed above. Postopera-
Apical vertebral body rotation (Nash-Moe) 3
T1 rib angle 5 tive patients complete Sections 1 and 2. This outcomes instru-
End instrumented vertebrae angulation 4 ment provides individual domain scores as well as an overall
Disc angulation 4 total score.
Classification of CSVL 5 Data were gathered for patients evaluated or treated for AIS
Total coronal and axial score 75
Lateral radiographs
from 1995 to 2000 at the seven scoliosis centers. A total of 620
T2–T12 (N ⫽ 20–45°) 5 patients from the Harms Study Group AIS database were avail-
T5–T12 (N ⫽ 20–40°) 5 able for review. Data for 355 of the patients were excluded
T2–T5 (10° normal) 5 because they did not meet the following inclusion criteria:
T12–L2 (N ⫽ ⫾10°) 5 11–18 years of age at diagnosis of AIS, curve magnitude ⱖ10°
Lordosis (N ⫽ ⫺40 to – 60°) 5
Total sagittal score 25 (diagnosis of scoliosis), radiographic and SRS outcomes data
recorded concurrently (measured ⬍1 month apart), and only
Deformity score 100 data recorded at latest surveillance visit or final preoperative
visit for presurgical data and latest postoperative visit (1, 2, 3,
or 4 years) for postsurgical data. These criteria result in no
of improved quality of life and reduced truncal more than two data sets per patient included for review. The
deformity.1,3,12,23 application of these inclusion criteria yielded 354 data sets
The Scoliosis Research Society (SRS) outcomes instru- from 265 patients available for review. Eighty-nine patients
had two data points: preoperative and latest postoperative. Of
ment is a disease-specific health-related quality-of-life
the 176 remaining patients with only one data point, 106 were
questionnaire that has been validated for use in the AIS preoperative data points and 70 were postoperative.
patient population.9 Although improvement in SRS out- Pearson’s correlation coefficient was calculated to identify
comes scores following surgery for AIS has been docu- associations between any of the radiographic measures, total
mented,3,12 there is little evidence of direct correlation of radiographic deformity scores, and the seven SRS domains as
the primary clinical measure of scoliosis (i.e., radio- well as the total SRS score. The SRS outcome domain scores
graphic Cobb angle) and the quality-of-life measures and total scores were normalized because the number of ques-
(i.e., SRS outcomes instrument scores). tions answered pretreatment differ from the number answered
Understanding that the radiographic process mea- posttreatment. The scores were normalized by dividing the
sures of patients with scoliosis reflect a component of the score by the number of questions answered. Because of the
disorder that is different from those measured by the SRS number of statistical comparisons being made, a P value of
⬍0.005 was considered statistically significant. Radiographic
outcomes instrument, it remains reasonable to assume
measures that were identified as being significantly correlated
that this disease-specific questionnaire would have some with SRS scores were then entered into a stepwise regression
correlation with the radiographic assessment of disease analysis. Linear regression analysis was also used to examine
severity. The purpose of this study was to test the validity whether gender, age, exposure to surgery, or severity of initial
of this assumption and determine whether any measures curve was predictive of SRS scores. Statistical measures were
in a standardized radiographic deformity system directly performed using Statistical Package for Social Science (Chica-
correlate with patient-reported quality of life as mea- go, IL).
sured by the SRS outcomes instrument.
Methods Results
A review of data collected on patients between 1995 and 2000 Of the 265 patients included for study, 36 (13.6%) were
at seven scoliosis centers within the United States provided male and 229 (86.4%) were female. Patient age ranged
material for study. All of the centers used a radiographic defor- from 11 to 21 years with a mean age of 14.7 years (SD
mity scoring system developed by the Harms Study Group to 2.5 years). Curve magnitude before treatment ranged
evaluate patients with AIS.3 This radiographic system assigns from 32° to 96° with a mean curve magnitude of 52.5°
point values to direct numerical measures taken from the stan- (SD 9.7°). Normalized SRS total score ranged from 1.32
dard posteroanterior and lateral upright scoliosis radiographs.
to 4.23 with a mean score of 2.8 (SD 0.43).
Therefore, standard coronal and sagittal Cobb angle measures
and radiographic measures of coronal and sagittal balance
A significant negative correlation, as indicated by
were available for study. The 15 radiographic measurements Pearson’s correlation coefficient, was found between the
and the respective points for the deformity scoring system are coronal measure of the thoracic scoliosis curve and the
presented in Table 1. Total Pain domain (r ⫽ ⫺0.22, P ⬍ 0.001), General
During the study collection period, centers began prospec- Self-Image domain (r ⫽ ⫺0.23, P ⬍ 0.001), General
tively using the SRS outcomes questionnaire as a quality-of-life Function domain (r ⫽ ⫺0.18, P ⬍ 0.003), and total SRS
2038 Spine • Volume 27 • Number 18 • 2002
Table 2. Correlations*
Total General General
Pain Self-Image Function
Thoracic curve ⫺0.22 ⫺0.23 ⫺0.18
Upper thoracic curve — — ⫺0.19
Lumbar curve ⫺0.20 ⫺0.23 — —
Coronal–axial 0.24 0.24 0.17
subscore
Sagittal subscore — — — —
Total deformity score 0.24 0.25 0.16
* All displayed r values are significant at the P ⬍ 0.005 level. Missing values indicate nonsignificant r values, which were all above the P ⬎ 0.05 level.
scores (r ⫽ ⫺0.22, P ⬍ 0.001). The coronal measure of
the lumbar scoliosis curve was found to be negatively
correlated with the Total Pain domain (r ⫽ ⫺0.20, P ⬍
0. 001), General Self-Image domain (r ⫽ ⫺0.23, P ⬍
0.001), and total SRS scores (r ⫽ ⫺0.26, P ⬍ 0.001). The
upper thoracic curve measure was found to be negatively
correlated with the General Function domain (r ⫽
⫺0.19, P ⫽ 0.001). As curve magnitude (Cobb angle)
increases, domain and total scores on the SRS tend to
decrease, hence the negative correlation. The coronal ra-
diographic deformity subscore as well as the total radio-
graphic deformity score were significantly positively cor-
related with the same SRS domain and total scores (r ⫽
0.1– 0.26, P ⬍ 0.005) (Table 2). These correlations are
positive because of the radiographic scoring system in
which higher scores are indicative of a smaller deformity.
As values on the radiographic score increase (indicating a
decreasing deformity), domain and total scores on the
SRS tend to increase as well. The remaining radiographic
parameters did not have a significant correlation with
any of the SRS scores.
Thus, the coronal radiographic measures of the tho-
racic curve, lumbar curve, and upper thoracic curve were
significantly correlated with selected SRS outcome mea-
sures. The Harms Study Group radiographic “coronal”
deformity subscore and the Harms Study Group radio-
graphic “total” deformity score were found to have sim-
ilar significant correlations. No sagittal radiographic
measures were significantly correlated with SRS out-
comes scores. Additionally, no radiographic measures
Figure 1. Thoracic curve as a predictor of pain scores (*total pain
score/number of pain questions answered); adjusted R2 ⫽ 0.05,
P ⬍ 0.001.
Postoperative Postoperative Total
Activity Self-Image Function Satisfaction SRS
— — — — ⫺0.22
— — — — —
— — — ⫺0.26
— — — — 0.26
— — — —
— — — — 0.26
taken in postoperative patients were correlated with the
postoperative domains of the SRS outcomes instrument
(Postoperative Self-Image, Postoperative Function, Ac-
tivity, and Satisfaction).
The radiographic measures that were found to have a
significant linear association with SRS scores were then
entered into a stepwise linear regression analysis as pre-
dictors of SRS scores. These predictors included thoracic,
upper thoracic, and lumbar curve magnitudes. The tho-
racic curve size was found to be the best predictor of
Total Pain domain scores (adjusted R2 ⫽ 0.05, P ⬍
0.001) (Figure 1). The upper thoracic curve was found to
be the best predictor of General Function domain scores
(adjusted R2 ⫽ 0.03, P ⬍ 0.001). The lumbar curve was
found to be the best predictor of General Self-Image do-
main scores (adjusted R2 ⫽ 0.05, P ⬍ 0.001) and total
SRS score (adjusted R2 ⫽ 0.07, P ⬍ 0.001) (Figure 2).
The radiographic deformity score was the next best pre-
dictor of total SRS score (adjusted R2 ⫽ 0.06, P ⬍ 0.001)
(Figure 3).
Neither age nor gender was found to be a significant
predictor of the three domains above (Total Pain, Gen-
eral Self-Image, General Function) or total SRS score, as
indicated by adjusted R2 values ⬍0.001 (P ⬎ 0.10).
When examined as an independent variable, a patient’s
exposure to surgery was found to predict the same or
more of the variance in SRS scores, as did the coronal
measures of lumbar curve, thoracic curve, and upper
thoracic curve. Exposure to surgery was a significant
predictor of improvement in the following domains: To-
Figure 2. Lumbar curve as a predictor of total SRS scores (*total
SRS score/number of questions answered); adjusted R2 ⫽ 0.07,
P ⬍ 0.001.
 Scoliosis Research Society Outcomes Instrument • Wilson et al 2039
Figure 3. Total radiographic deformity score as a predictor of total
SRS scores (*total SRS score/number of questions answered);
adjusted R2 ⫽ 0.06, P ⬍ 0.001.
tal Pain scores (adjusted R2 ⫽ 0.07, P ⬍ 0.001), General
Self-Image scores (adjusted R2 ⫽ 0.05, P ⬍ 0.001), and
total SRS scores (adjusted R2 ⫽ 0.06, P ⬍ 0.001). It was
not found to be a significant predictor of General Func-
tion scores (adjusted R2 ⫽ 0.002, P ⬎ 0.2).
Discussion
Early reviews examining outcomes of treatment for AIS
focused on patient-reported pain, largely without refer-
ence to curve magnitude. Edgar and Mehta6 as well as
Dickson et al4 reported that instrumentation and fusion
resulted in decreased reported pain compared with non-
operatively treated patients.
Subsequently, numerous questionnaires were used to
attempt to study the association of surgical treatment
and curve magnitude to patient-reported out-
come.5,17,18,21 In 1995, Haher et al10 performed a meta-
analysis of reports on 11,000 patients spanning 35 years
and concluded that curve correction was associated with
an increase in patient-reported satisfaction. These inves-
tigations pointed out the need for a standardized out-
comes questionnaire for use in the study of AIS treatment.
In 1999, Haher et al9 reported the results of a validated,
disease-specific outcomes instrument developed with the
support of the SRS for use in the study of AIS.
Recent investigations regarding the SRS outcomes in-
strument report an increase in scores following surgical
treatment for AIS; however, no significant correlation
with radiographic measures has been observed.3,12,23
The purpose of our current study was to further investi-
gate the association of radiographic measures and pa-
tient-reported outcomes. An attempt was made to exam-
ine whether radiographic parameters are directly
predictive of patient-reported quality-of-life measures
independent of the effect of surgery.
The results of this study indicate that the Cobb angle
measure of the coronal radiographic deformity is corre-
lated with reported Total Pain, General Self-Image, Gen-
eral Function, and total SRS outcome scores. In examin-
ing this association further, the relatively low R2 values
indicate that a greater Cobb angle is minimally predictive
of a lower SRS score (symptomatic) compared with a
smaller Cobb angle.
When examined as an independent predictor, surgical
status was also found to be a significant but weak pre-
dictor of Total Pain, General Self-Image, and total SRS
scores, with a general trend showing improvement in
scores after exposure to surgery. When examining the
relative R2 values (0.05– 0.07 for surgical status vs. 0.03–
0.07 for coronal radiographic measures), it can be seen
that the independent effect of undergoing surgery is as
powerful a predictor of SRS outcome domain scores as
that of coronal curve magnitudes. This finding would
support previous studies that report improvement in SRS
outcomes scores after surgery.
Whereas the Cobb angle measure of the coronal ra-
diographic deformity is correlated with reported Total
Pain, General Self-Image, General Function, and total
SRS outcome scores in this study, the relatively low R2
values indicate that there must be other variables that are
influencing the SRS outcomes scores. It is reasonable that
psychosocial factors may have a significant influence on
SRS scores. Issues regarding self-image, confidence, and
self-esteem may predominate over physical complaints.
These psychological issues are difficult to measure and may
or may not relate directly to a patient’s deformity.8,11,16
Recent literature does suggest that children and ado-
lescents can provide self-assessments that correlate with
physician assessment of outcomes.16 Parental reports on
two quality-of-life indexes, the Child Health Question-
naire and the American Academy of Orthopedic Sur-
geons Pediatric Outcomes Data Collection Instrument
(AAOS-PODCI), were found to correlate with a clini-
cian’s subjective rating of patient health.20 There were
several domains on each instrument that distinguished
between patients with different diagnoses, in particular
cerebral palsy and scoliosis. However, these correlations
were found to exist between functional outcome assess-
ments with no comparison to objective process mea-
sures. Additionally, outcomes literature suggests that
change occurring in relatively small amounts, as might
be expected concerning physical complaints in AIS, is
difficult to measure.7,19 All of these issues may be rele-
vant when examining the disparity between objective ra-
diographic process measures and patient-reported out-
come measures in AIS.
As a retrospective review, this study may be subject to
the sampling problems inherent to that mode of investi-
gation. Undoubtedly, some patients with AIS presenting
at the seven centers were not evaluated with the radio-
graphic deformity system or the SRS instrument. Many
patients were excluded from the study because of incom-
plete data, as the inclusion criteria required the radio-
graphs and SRS questionnaire to be obtained within 1
month’s time of each other. This issue of heterogeneity
within the sample may have been diminished by the large
sample size. The other issue concerns bias within the
sample. Because the questionnaires were not adminis-
tered within a standard protocol, patients may have been
biased by prior knowledge gained during previous con-
sultations. The effects of this potential bias within our
2040 Spine • Volume 27 • Number 18 • 2002
sample is difficult to address because we do not know
what knowledge of scoliosis the patients had when an-
swering these questionnaires. Despite these issues, we
think this retrospective review serves an important
purpose.
Conclusion
In this series of patients, the Cobb angle measure of the
major deformity has a statistically significant but weak
correlation with the reported total pain, general self-
image, and general function as measured by the SRS out-
comes instrument. These correlations are relatively low,
and none of the radiographic measures in this population
correlated with postoperative domain scores of the SRS
outcomes tool. Other studies have shown that the SRS
outcomes instrument may be of some use in assessing
individual patients before and after surgery.12 However,
the current study suggests that additional factors must
largely be affecting outcome scores aside from the radio-
graphic measures of deformity.
Key Points
● The Cobb angle measure of the major deformity
has a statistically significant correlation with the
reported total pain, general self-image, and gen-
eral function as measured by the SRS outcomes
instrument.
● This correlation, although statistically signifi-
cant, is weak.
● Postoperative radiographic measures did not
correlate significantly with the postoperative do-
mains of the SRS outcomes instrument.
Acknowledgments
The authors thank Tracey Gaynor, MA, Gail Huss, RN,
and Michelle Marks, PT, for their assistance with data
management and manuscript preparation and Depuy
AcroMed for their support of the Harms Study Group.
References
1. Betz RR, Shufflebarger H, McLain RF, et al. Anterior versus posterior instru-
mentation for the correction of thoracic idiopathic scoliosis. Spine 2001;26:
1095–100.
2. Collis DK, Ponseti IV. Long-term follow-up of patients with idiopathic sco-
liosis not treated surgically. J Bone Joint Surg Am 1969;51:425– 45.
3. D’Andrea LP, Betz RR, Lenke LG, et al. Do radiographic parameters corre-
late with clinical outcomes in adolescent idiopathic scoliosis? Spine 2000;23:
1795– 802.
4. Dickson JH, Erwin WD, Rossi D. Harrington instrumentation and arthro-
desis for idiopathic scoliosis. J Bone Joint Surg Am 1990;72:678 – 83.
5. Dickson JH, Mirkovic S, Noble PC, et al. Results of operative treatment of
idiopathic scoliosis in adults. J Bone Joint Surg Am 1995;77:513–23.
6. Edgar MA, Mehta MG. Long-term follow-up of fused and unfused idio-
pathic scoliosis. J Bone Joint Surg Br 1988;70:712– 6.
7. Fischer D, Stewart AL, Bloch DA, et al. Capturing the patient’s view of
change as a clinical outcome measure. JAMA 1999;282:1157– 62.
8. Guyatt GH, Jaeschke R. Measurements in clinical trials: choosing the appro-
priate approach. In: Spiker B, ed. Quality of Life Assessment in Clinical
Trials. New York, NY: Raven Press, 1990; 37– 46.
9. Haher TR, Gorup JM, Shin TM, et al. Results of the Scoliosis Research
Society instrument of evaluation of surgical outcome in adolescent idiopathic
scoliosis: a multi-center study of 244 patients. Spine 1999;24:1435– 40.
10. Haher TR, Merola A, Zipnick RI, et al. Meta-analysis of surgical outcome in
adolescent idiopathic scoliosis: a 35-year English literature review of 11,000
patients. Spine 1995;20:1575– 84.
11. Johnson J. Outcomes analysis in spinal research: how clinical research differs
from outcomes analysis. Orthop Clin North Am 1994;25:205–13.
12. Merola A, Brkaric M, Panagopoulos G, et al. A multi-center prospective
study of the surgical treatment of adolescent idiopathic scoliosis utilizing the
Scoliosis Research Society outcome instrument [abstract]. Scoliosis Research
Society, 35th Annual Meeting Program, Cairns, Australia, 2000.
13. Mielke CH, Lonstein JE, Denis F, et al. Surgical treatment of adolescent
idiopathic scoliosis. J Bone Joint Surg Am 1989;71:1170 –7.
14. Moe JH. A critical analysis of methods of fusion for scoliosis: an evaluation of
two hundred and sixty-six patients. J Bone Joint Surg Am 1958;40:529 –54.
15. Moskowitz A, Moe HH, Winter RB, et al. Long-term follow-up of scoliosis
fusion. J Bone Joint Surg Am 1980;62A:364 –76.
16. Pencharz J, Young NL, Owen JL, et al. Comparison of three outcomes
instruments in children. J Pediatr Orthop 2001;21:425–32.
17. Roye DP Jr, Farcy JP, Rickert JB, et al. Results of spinal instrumentation of
adolescent idiopathic scoliosis by King type. Spine 1992;17:270 –3.
18. Simmons ED Jr, Kowalski JM, Simmons EH. The results of surgical treat-
ment for adult scoliosis. Spine 1993;18:718 –24.
19. Taylor SJ, Taylor AE, Foy MA, et al. Responsiveness of common outcome
measures for patients with low back pain. Spine 1999;24:1805–12.
20. Vitale MG, Levy DE, Moskowitz AJ, et al. Capturing quality of life in
pediatric orthopedics: two recent measures compared. J Pediatr Orthop
2001;21:629 –35.
21. Waddell G, Reilly S, Trosney B, et al. Assessment of the outcome of low back
surgery. J Bone Joint Surg Br 1988;70:723–7.
22. Weinstein SL, Zavala DC, Ponseti IV. Idiopathic scoliosis: long term fol-
low-up and prognosis in untreated patients. J Bone Joint Surg Am 1981;63:
702–12.
23. White SF, Asher MA, Lai SM, et al. Patient’s perception of overall function,
pain, and appearance following primary posterior instrumentation and fu-
sion for idiopathic scoliosis. Spine 1999;24:1693–9.
Address reprint requests to
Peter O. Newton, MD
3030 Children’s Way, Suite 410
San Diego, CA 92123
E-mail: pnewton@chsd.org