Goal-determined metrics to assess outcomes

of exotropia surgery
Yoon-Hee Chang, MD, PhD,a Patrice Melvin, MPH,b and Linda R. Dagi, MDa
PURPOSE
METHODS

RESULTS

CONCLUSIONS

To present a goal-determined methodology for monitoring outcomes after surgery for

exotropia.
The goal-determined metric required surgeons to rank four possible goals preoperatively:
(1) binocular potential, (2) restoration of eye contact, (3) diplopia control; and (4) torticollis
management. Potential preoperative risk factors were noted. Goal-specific outcomes
criteria were applied to the latest sensory-motor examination, 2-6 months after surgery.
The medical records of patients who underwent surgery from 2007 to 2012 were retrospectively reviewed with respect to the goal-directed metric.
A total of 852 patients were evaluated in the study period: 411 for restoration of eye contact;
347 for binocular potential; 78 for diplopia resolution; and16 for torticollis management.
Excellent (62%) or good (16%) outcomes were achieved in 78%. Procedures to resolve
diplopia (OR, 6.56; 95% CI, 3.39-12.68) and to restore eye contact (OR, 3.74; 95% CI,
2.65-5.29) were more likely to result in excellent outcomes than procedures to improve
binocular potential. Simultaneous surgery for dissociated vertical deviation (OR, 0.38;
95% CI, 0.16-0.92) and preoperative near deviation $50D (OR, 0.27; 95% CI, 0.170.42) limited likelihood of an excellent outcome. Outcomes monitored by simultaneous
rather than alternate prism and cover test were more likely graded excellent (OR, 5.16;
95% CI, 3.50-7.62). Applying motor criteria from the binocular potential goal to the entire
cohort diminished putative outcomes (P \ 0.001).
Goal-determined metric monitoring outcomes of exotropia surgery provides outcomes
germane to the reason for intervention, enables analysis of risk factors affecting outcomes,
and facilitates reporting on heterogeneous populations. ( J AAPOS 2015;19:304-310)

xotropia may be variably expressed and associated

with conditions as distinct as intermittent exotropia with preservation of fusion, congenital exotropia absent fusion, overcorrection after surgery for
esotropia, Duane syndrome, orbital trauma, craniofacial
dysgenesis, oculomotor nerve palsy, and paralytic pontine
exotropia. Even for patients with the same underlying
cause, the goal of surgical treatment may differ based on
symptoms and social needs.1 Remediation of intermittent
exotropia for the purpose of improving binocularity demands, at a minimum, good control of motor alignment.

Author affiliations: aDepartment of Ophthalmology, Boston Childrens Hospital, Harvard

Medical School, Boston, Massachusetts; bProgram for Patient Safety & Quality, Boston
Childrens Hospital, Boston, Massachusetts
This research was supported, in part, by endowment funds from the Childrens Hospital
Ophthalmology Foundation Chair, held by Linda R Dagi, MD.
Presented as a poster at the 40th Annual Meeting of the American Association for
Pediatric Ophthalmology and Strabismus, Rancho Mirage, California, April 2-6, 2014.
Submitted February 13, 2015.
Revision accepted April 21, 2015.
Published online July 30, 2015.
Correspondence: Dr. Linda R. Dagi, MD, Boston Childrens Hospital, 300 Longwood
Avenue, Boston, MA 02115 (email: linda.dagi@childrens.harvard.edu).
Copyright 2015 by the American Association for Pediatric Ophthalmology and
Strabismus.
1091-8531/$36.00
http://dx.doi.org/10.1016/j.jaapos.2015.04.009

304

However, for the patient having surgery to eliminate

diplopia, excellent postoperative motor alignment is no solace if diplopia persists. In contrast, for patients whose primary goal is restoration of eye contact, surgery is best
aimed at postoperative centration of the pupillary light reflex rather than the results of a prism cover test, especially
in the presence of craniofacial abnormalities, or a positive
angle kappa in the absence of fusion.2 In addition, for patients with exotropia and torticollis from Duane syndrome,
outcomes may be best monitored on the basis of successful
remediation of torticollis.
Responding to the increased interest in metrics suitable
for monitoring surgical outcomes,3,4 we developed a goaldetermined outcomes methodology to appraise the results
of strabismus surgery in large and heterogeneous populations. In 2014 we reported the results of the successful
application of a goal-determined outcomes metric developed to appraise the surgical treatment of all forms of esotropia.5 The purpose of this study is to report surgical
outcomes based on application of a similar metric designed
for surgical treatment of exotropia. Goal-determined outcomes are compared to those that would have been reported with indiscriminate application of standard motor
criteria regardless of the primary goal. In addition, we
used the metric to assess the effect of pre- and perioperative
risk factors, demographics, and the specific methodology

Journal of AAPOS

Volume 19 Number 4 / August 2015

Chang, Melvin, and Dagi

305

Table 1. Goal-determined outcome criteria for exotropia surgery

Goal

Result

Criteria

Excellent

Motor: XT \ 10 PD or ET \ 6 PD at D and N by SPCTa

Sensoryb: Near stereoacuity \2 octaves worse than preoperative status and
not diminished to nilc; if stereoacuity not available, no loss of fusion by W4D
Motor: 10 PD # XT \ 15 PD or 6 PD # ET # 10 PD at D and N by SPCTa
Sensoryb: Near stereoacuity \2 octaves worse than preoperative status and
not diminished to nil; If stereoacuity not available, no loss of fusion by W4D
Motor: XT $ 15 PD or ET . 10 PD at D or N by SPCTa or planned reoperation
for horizontal deviation
Sensoryb: Near stereoacuity $2 octaves worse than preoperative status or
diminished to nil; If stereoacuity not available, loss of fusion by W4D

Binocular potential

Good
Poor

Restoration of eye contact (reconstructive)

Excellent
Good
Poor

XT or ET # 10 PDd,e
10 \ XT or ET # 15 PDd,e
XT or ET . 15 PDd,e or planned reoperation for horizontal deviation

Excellent
Good
Poor

Torticollisf #8
8 \ torticollis # 12
Torticollis . 12 or planned reoperation for torticollis

Excellent
Good
Poor

No diplopia in primaryg,h,i
Diplopia controlled with the aid of a prism#10 PDg,h,i in primary
Diplopia requiring occlusion or planned reoperation for diplopia

Resolution of torticollis

Resolution of diplopia

APCT, alternate prism cover test; D, distance; ET, esotropia; N, near; PD, prism diopter; SPCT, simultaneous prism cover test; XT, exotropia; W4D,
Worth 4-dot test.
a
Order of preference for angle used: SPCT . APCT . Krimsky.
b
Accept W4D fusion if stereoacuity data not available.
c
One octave represents doubling of the stereoacuity (eg from 100 to 200 arcsec).6
d
Order of preference for angle used: Krimsky . SPCT . APCT.
e
Alignment at near unless stated goal is distance.
f
Torticollis as assessed at distance unless stated goal at near.
g
May have occasional diplopia in primary but controlled without prism correction.
h
May have diplopia away from primary.
i
Preexisting vertical deviation controlled with prism does not affect outcome if no increase in prism correction needed.

of monitoring postoperative motor alignment on reported

surgical outcomes.

Methods
The Boston Childrens Hospital Center for Clinical Investigation
provided approval for this study. In 2005 the Program for Patient
Safety and Quality in collaboration with the Department of
Ophthalmology at Boston Childrens Hospital initiated a quality
improvement project piloting a new methodology to monitor outcomes of surgery for exotropia. A form was created for surgeons to
prospectively document details of the oculomotor examination,
known risk factors likely to influence outcomes, demographic
data, and surgery performed, in addition to a ranked order of the
surgeons goals for performing surgery. The potential preoperative
risk factors specifically documented were bilateral vision limitation
(eg, albinism), prior strabismus surgery, prior surgery for retinal
detachment, antecedent orbital trauma, conditions resulting in hyper- or hypotonia, craniosynostosis or craniofacial anomalies, oculomotor nerve palsy, trochlear nerve palsy, Duane syndrome,
Graves orbitopathy, and congenital fibrosis of the extraocular
muscles (CFEOM). Age at time of surgery, preoperative alignment
at near and distance and surgical details, specifically, simultaneous
surgery for A and V patterns, nystagmus, dissociated vertical devi-

Journal of AAPOS

ation, other vertical deviations, and the use of adjustable sutures

were documented to enable investigation of effect on outcome.
Four goals were ranked: (1) binocular potential, for patients
believed to have the potential to develop or maintain binocular
vision; (2) reconstructive, to establish the appearance of normal
eye alignment and eye contact; (3) diplopia resolution; and (4)
reduction of torticollis, for patients with a compensatory head
posture.
Ocular alignment with refractive correction was recorded at
distance (20 feet) and near (13 inches) by alternate prism and cover
test (APCT), simultaneous prism and cover test (SPCT), or Krimsky. Worth 4-Dot and stereoacuity were measured at distance and
near for cooperative patients. Near stereoacuity was measured using Titmus stereo test (Stereo Optical, Chicago, IL), Randot stereotest (Stereo Optical, Chicago, IL), or Randot Preschool
Stereoacuity Test (Stereo Optical, Chicago, IL), and Vectographic Project-o-Chart slide (American Optical, Southbridge,
MA) was used to test distant stereoacuity. A change of 2 octaves
or more was considered to exceed testretest variability.6 One
octave represents the doubling of stereoacuity (eg, from 100 to
200 arcsec) and , accordingly, a 2-octave decline would diminish
stereoacuity from 100 to 400 arcsec or from 40 to 160 arcsec. A
complete sensorimotor evaluation, presence of diplopia in primary position at distance and at near, and measurement of

306

Chang, Melvin, and Dagi

Volume 19 Number 4 / August 2015

Table 2. Demographic and clinical features of cohort by primary goal

Binocular potential
(N 5 347)

Reconstructive
(N 5 411)

Resolution of
Diplopia (N 5 78)

Resolution of
Torticollis (N 5 16)

Total (N 5 852)

48.0 (34.7-61.5)

9.2 (6.7-28.1)

11.7 (5.6-38.2)

Age years, median (IQR)

5.9 (4.4-9.0)
22.8 (9.6-44.7)
Age group, n (%)
0-2 years
14 (4.0)
3 (0.7)
3-5 years
108 (31.1)
36 (8.8)
6-11 years
173 (49.9)
86 (20.9)
$12 years
52 (15.0)
286 (69.6)
Median deviation, PD (IQR)
Distance
30 (20-35)
30 (25-35)
Near
25 (20-35)
35 (25-49)
X(T), n (%)
284 (81.8%)
85 (20.9%)
Adjustable sutures, n (%)
59 (17.1%)
252 (61.3%)
Simultaneous cyclovertical muscle surgery, n (%)a
A pattern
12 (3.5)
36 (8.8)
V pattern
101 (29.1)
58 (14.1)
DVD
4 (1.2)
16 (3.9)
Other
14 (4.0)
49 (11.9)
Totalb
125 (36.1%)
147 (35.8%)

0
0
4 (5.1)
74 (94.9)

0
2 (12.5)
8 (50.0)
6 (37.5)

17 (2.0)
146 (17.1)
271 (31.8)
418 (49.1)

20 (12-30)
25 (14-40)

20 (17-25)
25 (19-30)

30 (20-40)
30 (20-40)

32 (41.0%)

5 (31.3%)

406 (47.8%)

72 (92.3%)

7 (43.8%)

390 (45.8%)

4 (5.1)
9 (11.5)
3 (3.9)
22 (28.2)
35 (44.9%)

1 (6.3)
4 (25.0)
1 (6.25)
5 (31.3)
10 (62.5%)

53 (6.2)
172 (20.2)
24 (2.8)
90 (10.6)
317 (37.2%)

DVD, dissociated vertical deviation; IQR, interquartile range; PD, prism diopters; X(T), intermittent exotropia.
a
One or more simultaneous cyclovertical muscle procedures.
b
Some patients have more than one type of vertical deviation.
torticollis by goniometer performed 2-6 months after surgery
were used to appraise outcome. The latest complete examination
was used when more than one examination took place within this
period.
Goal-determined outcome criteria were applied based on the
primary goal, and the outcome was designated as excellent,
good, or poor (Table 1). When the designation of primary goal
had not been submitted prior to the first postoperative evaluation
(forms missing or incomplete), the quality assurance team conducted a determination of the surgical goal based on the patients
preoperative clinical history, subjective complaint, and clinical examination. This occurred in less than 5% of our cases. For those
with diplopia or with documented torticollis, remediation of these
was ranked, based on symptoms. For those below 12 years of age
with apparent binocular potential this became the primary goal.
For older patients, or for younger patients without evidence of
any binocularity or with poor acuity, the reconstructive goal
was considered primary. Patients were excluded if there was no
complete postoperative sensorimotor examination within the 26 month window (allowed window, 1.2-7 months).
Primary outcome was success of surgical intervention (excellent, good, or poor) at 2-6 months by criteria for the primary
goal. We then compared outcomes measured by this metric to
those that would have been recorded based on indiscriminate
application of our binocular motor criteria to all patients. Motor
criteria from the binocular potential group was chosen based on a
review of existing literature.7 Secondary analysis determined
which preoperative risk factors or demographic features
decreased the probability of an excellent outcome. In addition,
we studied the effect of simultaneous surgery for alphabet patterns, other vertical deviations, and null point nystagmus, and
whether performing surgery with adjustable sutures might affect
outcome. Finally, we appraised the effect that methodology cho-

sen to document postoperative motor alignment (SPCT vs APCT

vs Krimsky) had on reported outcomes.

Statistical Analysis
Descriptive analyses of patient pre and perioperative characteristics
across objective groups (binocular, diplopia, reconstructive, and
torticollis) were summarized using medians (interquartile range)
for continuous variables and proportions for categorical variables.
Patient risk factors were presented as an overall binary category
(yes/no) and, because a patient may have more than one risk factor,
individually by risk factor type. Surgical outcomes, categorized as
excellent, good, and poor, were presented overall, and by surgical
objective. Our univariate analysis assessed the association of individual categorical factors on surgical outcomes using the c2 test
and Fisher exact test where appropriate. We assessed differences
in median patient age as well as in near and distance deviation at
the time of surgical intervention across surgical outcome groups using the Kruskal-Wallis test. Because the categories of surgical outcomes in our study were ordinal in nature yet the distance between
excellent, good, and poor outcomes could not be assumed, we used
an ordered logistic regression model for our multivariable analysis.
Covariates with a P value of #0.2 in the univariate analysis were
included in the ordered logistic regression analysis to identify predictors of surgical outcome. Bowkers test of symmetry was used to
compare differences in outcomes using goal determined versus single motor criteria of binocular potential methods. All analyses were
performed using SAS version 9.4 (SAS Institute, Cary, NC).

Results
From 2007 to 2012, 909 procedures were performed for
the treatment of exotropia alone, or as part of a more complex strabismus repair. Of these, follow-up sensorimotor

Journal of AAPOS

Volume 19 Number 4 / August 2015

FIG 1. Outcomes of surgical treatment of exotropia overall and by primary goal. Outcomes differ significantly by subgroup (P \ 0.001).

evaluation at 2-6 months was available for 852 surgeries.

The mean time to included postoperative outcome examination was 2.80  1.4 months.
Table 2 summarizes the demographic and clinical features across the four primary goal groups. For 411 cases
(48%) the primary goal was reconstructive; for 347
(41%), binocular potential. Of the 852 cases, 440 (52%)
had associated risk factors; prior strabismus surgery was
the most common.
Applying the goal-specific tool to the entire cohort, 528
(62%) had excellent outcomes, and 136 (16%) had good
outcomes. Best outcomes were achieved for surgery to control diplopia (82% excellent, 12% good), with more
modest results for surgery to improve eye contact (69%
excellent and 13% good) and resolve torticollis (63% excellent and 13% good). See Figure 1.
We compared our outcomes utilizing this goal-specific
metric with outcomes for the same cohort assessed by standard motor criteria and found the putative outcomes declined
significantly using the latter (P \ 0.001). See Figure 2.
Univariate Analysis
Univariate analysis of the effect of each risk factor or demographic feature demonstrated a statistically significant
reduction in excellent outcomes for those with a preoperative angle $50D at near (P 5 0.007) and at distance (P \
0.001). There was a greater risk of poor outcome for those
having surgery before 24 months of age (P \ 0.001). Patients with CFEOM, Duane syndrome, and prior surgery
for retinal detachment had fewer excellent outcomes, but
these groups were too small for adequate statistical analysis
(data not shown).
Surgery with adjustable sutures (P 5 0.002) increased
the likelihood of an excellent outcome. Outcomes based
on SPCT rather than APCT monitoring were associated
with more frequent (documentation of) excellent outcomes
(Table 3).

Journal of AAPOS

Chang, Melvin, and Dagi

307

FIG 2. Comparison of outcomes by the goal-determined metric (GD)

and by single motor criteria of binocular potential goal (BP) in entire
patients and goal-specific subgroups.

Multivariable Analysis
Multivariable analysis confirmed that the outcomes
differed by the primary goal of surgery (P \ 0.001). Procedures performed primarily to resolve diplopia (OR, 6.56;
95% CI, 3.39-12.68) or to restore eye contact (OR, 3.74;
95% CI, 2.65-5.29) were more likely to result in an excellent outcome than those designed to improve binocular potential. Patients with a preoperative deviation of $50D at
near (OR, 0.27; 95% CI, 0.17-0.42) and those with simultaneous surgery for dissociated vertical deviation (OR,
0.38; 95% CI, 0.16-0.92) were at higher risk for poor surgical outcomes. The independent positive effect of
monitoring outcome by SPCT rather than APCT or
Krimsky was confirmed (Table 4).

Discussion
Since the 1960s8 an immense body of literature has been
published on outcomes of surgery for exotropia. Most
studies are noncomparable, because they describe outcomes for patients of differing ages, etiologies, varied inclusion and exclusion criteria, different follow-up
periods, and, most importantly, differing outcomes criteria
and methodology for monitoring outcome.9-16 A general
lack of consensus in defining satisfactory outcomes
further complicates the interpretation of results.
Consequently, reported success rates vary from 41% to
81%.9-16 According to a recent review,7 56 studies were
published on outcomes of surgery for intermittent exotropia alone in the last 10 years, and outcome measures variously included ocular alignment, stereopsis, control,
visual acuity, reoperation rate, and postoperative drift.
Even for motor alignment, there were 11 different definitions of successful alignment, including various methodologies for monitoring alignment (SPCT vs APCT) and
degree of postoperative misalignment deemed satisfactory.
This study presents a goal-specific methodology, designed

308

Chang, Melvin, and Dagi

Volume 19 Number 4 / August 2015

Table 3. Univariate analysis: characteristics impacting surgical outcomes

Poor (N 5 188)
Primary goal, n (%)
Diplopia
5 (2.7)
Reconstructive
76 (40.4)
Binocular potential
103 (54.8)
Torticollis
4 (2.1)
Age at surgery
Median (IQR)
8.6 (4.9-23.3)
Age group, n (%)
0-2 years
9 (4.8)
3-5 years
40 (21.5)
6-11 years
66 (35.5)
$12 years
71 (38.2)
Deviation at distance
Median, PD (IQR)
30 (22-45)
Deviation group, n (%)
#25 PD
66 (35.5)
26-35 PD
53 (28.5)
36-49 PD
27 (14.5)
$50 PD
40 (21.5)
Deviation at near
Median, PD (IQR)
35 (25-50)
Deviation group, n (%)
#25 PD
70 (37.4)
26-35 PD
43 (23.0)
36-49 PD
27 (14.4)
$50 PD
47 (25.1)
Motor examination, n (%)
APCT
143 (79.4)
Krimsky
18 (10.0)
SPCT
19 (10.6)
Any risk factor, n (%)
No
85 (45.2)
Yes
103 (54.8)
Adjustable sutures, n (%)
No
116 (61.7)
Yes
72 (38.3)
Simultaneous surgery for DVD, n (%)
No
180 (95.7)
Yes
8 (4.3)

Good (N 5 136)

Excellent (N 5 528)

P value
\0.001

9 (6.6)
52 (38.2)
73 (53.7)
2 (1.5)

64 (12.1)
283 (53.6)
171 (32.4)
10 (1.9)

9.0 (5.5-35.9)

14.3 (6.4-42.1)

1 (0.7)
25 (18.4)
53 (39.0)
57 (41.9)

7 (1.3)
79 (15.0)
150 (28.5)
290 (55.1)

25 (20-40)
70 (52.2)
24 (17.9)
19 (14.2)
21 (15.7)
30 (20-40)

30 (20-35)

\0.001
\0.001

0.006
\0.001

234 (44.8)
167 (32.0)
70 (13.4)
51 (9.8)
30 (20-40)

52 (38.8)
36 (26.9)
23 (17.2)
23 (17.2)

231 (44.2)
136 (26.0)
92 (17.6)
64 (12.2)

97 (71.9)
13 (9.6)
25 (18.5)

264 (50.9)
53 (10.2)
202 (38.9)

68 (50.0)
68 (50.0)

259 (49.1)
269 (51.0)

85 (62.5)
51 (37.5)

261 (49.4)
267 (50.6)

136 (100)
0

512 (97.0)
16 (3.0)

0.020
0.007

\0.001

0.608
0.002
0.066

APCT, alternate prism cover test; DVD, dissociated vertical deviation; IQR, interquartile range; PD, prism diopter; SPCT, simultaneous prism and
cover test.

after review of pertinent literature,17-21 that enables

evaluation of an entirely heterogeneous population
treated with an array of surgical approachessome with
simultaneous surgery for other deviations, and others
with known preoperative risk factors. Failure to return
for postoperative evaluation within the required time
frame was the only reason for exclusion.
For our binocular potential group we chose alignment
within 10D of orthophoria, because this metric has been
associated with improved sensory function.17 These motor
criteria were also used to compare outcomes generated by
our methodology to those that would have been reported
on the basis of indiscriminant application of motor alignment to the entire cohort. In addition to motor criteria,
we included a sensory measure for our binocular potential
subgroup18,22 and assessment of diplopia control for the
diplopia subgroup.

Motor alignment for the reconstructive subgroup prioritizes Krimsky results, based on our clinical experience.
Outcomes criteria for those seeking remediation of torticollis were based on indications published by Barbe and
colleagues.23 Because postoperative drift toward exodeviation begins to stabilize after the sixth postoperative week
and alignment becomes fairly constant by 6 months,24 we
analyzed the outcome at the latest follow-up in the 2-6
month window to allow for resolution of cases of transient
postoperative esotropia and diplopia.
Outcomes differed significantly by the primary goal of
surgery. In our study, 82% of patients having surgery to
resolve diplopia had an excellent outcome and were 6.56
times more likely to have an excellent outcome than those
having surgery for binocular potential. Resolution of
diplopia improves quality of life and payers may be interested in the associated reduced risk of falling.25 Procedures

Journal of AAPOS

Volume 19 Number 4 / August 2015

Chang, Melvin, and Dagi

Table 4. Multivariable analysis. Ordered logistic regression model:

Characteristics affecting odds ratio of achieving excellent outcome
for surgical management of exotropia
Multivariate model
Characteristic
Primary goal
Diplopia
Reconstructive
Torticollis
Binocular potential
Simultaneous surgery for DVD
Yes
No
Near deviation
26-35 PD
36-49 PD
$60 PD
#25 PD
Motor examination
SPCT
Krimsky
APCT

OR

95% CI

6.56
3.74
2.96
1.00

(3.39-12.68)
(2.65-5.29)
(0.97-9.04)

0.38
1.00

(0.16-0.92)

0.89
0.84
0.27
1.00

(0.61-1.29)
(0.54-1.32)
(0.17-0.42)

5.16
1.35
1.00

(3.50-7.62)
(0.82-2.23)

P value
\0.001

0.033
\0.001

\0.001

APCT, alternate prism and cover test; CI, confidence interval; OR, odds
ratio; DVD, dissociated vertical deviation; PD, prism diopter; SPCT,
simultaneous prism and cover test.

restoring eye contact were 3.74 more likely to have excellent outcomes than those for binocular potential. The psychosocial benefits of restoring normal eye contact and its
importance for educational and employment options have
been well described.26-30
Hatt and colleagues21 recently compared outcome
criteria performance in adult strabismus surgery, evaluating the performance of motor, diplopia, and healthrelated quality of life criteria alone and in combination
for diplopic, nondiplopic, or atypical diplopic patients.
They found that motor criteria do not fully represent the
patients postoperative status, although applying motor
criteria alone yielded the highest success rate. They
concluded that combining diplopia criteria with motor
criteria provided a more clinically relevant standard for
judging the success of adult strabismus surgery. Although
excellent criteria in the adult population, they cannot be
readily applied to many younger patients, including the
majority of those included in our study.
The modest success for procedures designed to enhance
binocular potential is consistent with prior reports on the
treatment of patients with intermittent exotropia; 82% of
our patients having surgery to improve binocular potential
had intermittent exotropia, and this is historically a challenging population to treat. Several factors contribute to
this challenge. The first is the variability in control of strabismus seen in this subpopulation. Although our metric favors application of motor criteria based on SPCT rather
than APCT, SPCT data was not available on the majority
(64%) of this subgroup. As both measures were, unfortunately, not documented on most patients, we could not,
without introducing further bias, report outcomes based
on one subgroup or the other. This is a limitation of this

Journal of AAPOS

309

study, and of the metric, as it was initially implemented.

Most recently, all providers in our hospital have been
encouraged to document both SPCT and APCT results
to enable more sophisticated analysis of outcomes and
comparison to other studies.
It is noteworthy that the reported odds ratio favoring
excellent outcomes for those graded by SPCT was based
on multivariable analysis and thus independent of primary
goal and risk factors. This confirms that, regardless of the
methodology used to monitor postoperative motor alignment, binocular goals are harder to achieve. Prior publications31-33 have established that variable fusion and control
of alignment in patients with intermittent exotropia, the
dominant subpopulation, remains an outstanding clinical
concern and reoperation is common.
In addition to the methodology for monitoring motor
alignment (SPCT vs APCT), fewer excellent outcomes
for those seeking binocular goals may result from the narrower range of motor alignment, both at distance and at
near, and the mandated sensory criteria for success.
Multivariable analysis demonstrated that simultaneous
surgery for dissociated vertical deviation diminished outcomes. This is consistent with the findings of PrattJohnson and colleagues,9 who reported that the presence
of even a small vertical deviation decreased the chance of
successful exotropia surgery. Oh and Hwang,14 on the
other hand, failed to find preoperative characteristics
affecting success. Unlike some previous studies,11,14-16 we
demonstrated that very large preoperative deviation (at
near) diminished the likelihood of success. Some
surgeons hesitate to operate on more than 2 horizontal
muscles simultaneously, and a larger preoperative
deviation may be more difficult to accurately measure.
We note several limitations to the goal-determined
metric as originally conceived. The first is the difference
in outcomes when evaluating subpopulations based on
SPCT versus APCT. Remediation of this limitation is
already underway with more recent documentation of
both SPCT and APCT results on all patients. In addition,
outcomes reported are based on goals ranked by a variety
of surgeons, whose perspectives may vary. One may
choose binocular criteria for a young teen with intermittent exotropia and retained fusion, whereas another might
prioritize the reconstruction goal. Of the 103 patients
who had a poor outcome in the binocular subgroup, 6 patients were ranked as poor solely due to sensory loss
rather than undesirable motor alignment. If the reconstructive goal had been chosen for those patients, the
outcome would have been excellent. We are currently
developing a modified outcomes metric that considers
all ranked goals in a weighted manner.
The goal-determined outcomes metric facilitates reporting on outcomes to patients, providers and payers
based on goals germane to the reason for surgery.
Assuming patients return for follow-up examination,
none are excluded for analysis. The methodology enables
recognition of goal groups and risk factors that may

310

Chang, Melvin, and Dagi

positively or negatively impact outcomes and it is readily

applicable, even in settings with highly heterogeneous populations and surgeons who favor different surgical techniques.
References
1. Granet DB. Strabismus: aligning the doctors vision with the patients
need. Br J Ophthalmol 2011;95:443-4.
2. Hunter DG. Benefits of strabismus surgery in patients with one blind
eye. Arch Ophthalmol 1995;113:404.
3. Beauchamp CL, Beauchamp GR, Stager DR Sr, Brown MM,
Brown GC, Felius J. The cost utility of strabismus surgery in adults.
J AAPOS 2006;10:394-9.
4. Kushner BJ. The efficacy of strabismus surgery in adults: a review for
primary care physicians. Postgrad Med J 2011;87:269-73.
5. Ehrenberg M, Nihalani BR, Melvin P, Cain CE, Hunter DG,
Dagi LR. Goal-determined metrics to assess outcomes of esotropia
surgery. J AAPOS 2014;18:211-16.
6. Adams WE, Leske DA, Hatt SR, Holmes JM. Defining real change in
measures of stereoacuity. Ophthalmology 2009;116:281-5.
7. Chiu AK, Din N, Ali N. Standardising reported outcomes of surgery
for intermittent exotropiaa systematic literature review. Strabismus
2014;22:32-6.
8. Burian HM, Spivey BE. The surgical management of exodeviations.
Trans Am Ophthalmol Soc 1964;62:276-306.
9. Pratt-Johnson JA, Barlow JM, Tillson G. Early surgery in intermittent exotropia. Am J Ophthalmol 1977;84:689-94.
10. Richard JM, Parks MM. Intermittent exotropia. Surgical results in
different age groups. Ophthalmology 1983;90:1172-7.
11. Stoller SH, Simon JW, Lininger LL. Bilateral lateral rectus recession
for exotropia: a survival analysis. J Pediatr Ophthalmol Strabismus
1994;31:89-92.
12. Kushner BJ. Selective surgery for intermittent exotropia based on distance/near differences. Arch Ophthalmol 1998;116:324-8.
13. Maruo T, Kubota N, Sakaue T, Usui C. Intermittent exotropia surgery in children: long term outcome regarding changes in binocular
alignment. A study of 666 cases. Binocul Vis Strabismus Q 2001;16:
265-70.
14. Oh JY, Hwang JM. Survival analysis of 365 patients with exotropia after surgery. Eye (Lond) 2006;20:1268-72.
15. Ekdawi NS, Nusz KJ, Diehl NN, Mohney BG. Postoperative outcomes in children with intermittent exotropia from a populationbased cohort. J AAPOS 2009;13:4-7.
16. Chia A, Seenyen L, Long QB. Surgical experiences with two-muscle
surgery for the treatment of intermittent exotropia. J AAPOS 2006;
10:206-11.

Volume 19 Number 4 / August 2015

17. Kraft SP. Outcome criteria in strabismus surgery. Can J Ophthalmol
1998;33:237-9.
18. Pineles SL, Ela-Dalman N, Zvansky AG, Yu F, Rosenbaum AL.
Long-term results of the surgical management of intermittent exotropia. J AAPOS 2010;14:298-304.
19. Reinecke RD, Sterling R, Wizov S. Accuracy of judgments of the
presence or absence of eccentric (non-primary) gaze and the presence
or absence of strabismus. Binocul Vis Strabismus Q 1991;6:189-96.
20. Larson SA, Keech RV, Verdick RE. The threshold for the detection of
strabismus. J AAPOS 2003;7:418-22.
21. Hatt SR, Leske DA, Liebermann L, Holmes JM. Comparing outcome
criteria performance in adult strabismus surgery. Ophthalmology
2012;119:1930-36.
22. Buck D, Powell CJ, Sloper JJ, Taylor R, Tiffin P, Clarke MP. Surgical
intervention in childhood intermittent exotropia: current practice and
clinical outcomes from an observational cohort study. Br J Ophthalmol 2012;96:1291-5.
23. Barbe ME, Scott WE, Kutschke PJ. A simplified approach to the
treatment of Duanes syndrome. Br J Ophthalmol 2004;88:131-8.
24. Yang HK, Hwang JM. Bilateral vs unilateral medial rectus resection
for recurrent exotropia after bilateral lateral rectus recession. Am J
Ophthalmol 2009;148:459-65.
25. Merrill K, Satterfield D, OHara M. Strabismus surgery on the elderly
and the effects on disability. J AAPOS 2010;14:196-8.
26. Coats DK, Paysse EA, Towler AJ, Dipboye RL. Impact of large angle
horizontal strabismus on ability to obtain employment. Ophthalmology 2000;107:402-5.
27. Olitsky SE, Sudesh S, Graziano A, Hamblen J, Brooks SE, Shaha SH.
The negative psychosocial impact of strabismus in adults. J AAPOS
1999;3:209-11.
28. Nelson BA, Gunton KB, Lasker JN, Nelson LB, Drohan LA. The
psychosocial aspects of strabismus in teenagers and adults and the
impact of surgical correction. J AAPOS 2008;12:72-76.e1.
29. Hatt SR, Leske DA, Kirgis PA, Bradley EA, Holmes JM. The effects
of strabismus on quality of life in adults. Am J Ophthalmol 2007;144:
643-7.
30. Paysse EA, Steele EA, McCreery KM, Wilhelmus KR, Coats DK. Age
of the emergence of negative attitudes toward strabismus. J AAPOS
2001;5:361-6.
31. Hatt SR, Mohney BG, Leske DA, Holmes JM. Variability of control
in intermittent exotropia. Ophthalmology 2008;115:371-376.e2.
32. Buck D, Clarke MP, Haggerty H, et al. Grading the severity of intermittent distance exotropia: the revised Newcastle Control Score. Br J
Ophthalmol 2008;92:577.
33. Hatt SR, Leske DA, Liebermann L, et al. Associations between
health-related quality of life and the decision to perform surgery for
childhood intermittent exotropia. Ophthalmology 2014;121:883-8.

Seeing Is Believing
Arnall Patz would be disappointed to see that many ophthalmologists avoid treating
retinopathy of prematurity today. I dont think he would be angry; he wasnt the type to
get angry. But he might shake his head a little, and perhaps scratch it out of confusion,
because the unselfish man from rural Georgia who spent his own money no, borrowed
the money to conduct the first sight-saving study in premature infants would simply
not understand the way we think today.
We dont want to treat babies because were afraid of getting sued.
Andrew Lam, Saving Sight (Bokeelia, Fla.: Irie Books, 2013), 107.

Journal of AAPOS