Research

Original Investigation

Effect of a Teleretinal Screening Program

on Eye Care Use and Resources
Joel E. Chasan, MD; Bill Delaune, PhD; April Y. Maa, MD; Mary G. Lynch, MD

Author Audio Interview at

IMPORTANCE Telemedicine is a useful clinical method to extend health care to patients with jamaophthalmology.com
limited access. Minimal information exists on the subsequent effect of telemedicine activities
on eye care resources.

OBJECTIVE To evaluate the effect of a community-based diabetic teleretinal screening

program on eye care use and resources.

DESIGN, SETTING, AND PARTICIPANTS The current study was a retrospective medical record
review of patients who underwent diabetic teleretinal screening in the community-based
clinics of the Atlanta Veterans Affairs Medical Center from October 1, 2008, through March
31, 2009, and who were referred for an ophthalmic examination in the eye clinic.

EXPOSURES Clinical medical records were reviewed for a 2-year period after patients were
referred from teleretinal screening. The following information was collected for analysis:
patient demographics, referral and confirmatory diagnoses, ophthalmology clinic visits,
diagnostic procedures, surgical procedures, medications, and spectacle prescriptions.

MAIN OUTCOMES AND MEASURES The accuracy between referring and final diagnoses and the
eye care resources that were used in the care of referred patients.

RESULTS The most common referral diagnoses were nonmacular diabetic retinopathy
(43.2%), nerve-related disease (30.8%), lens or media opacity (19.1%), age-related macular
degeneration (12.9%), and diabetic macular edema (5.6%). The percentage of agreement
Author Affiliations: Department of
among these 5 visually significant diagnoses was 90.4%, with a total sensitivity of 73.6%. Ophthalmology, Emory University
Diabetic macular edema required the greatest number of ophthalmology clinic visits, School of Medicine, Atlanta, Georgia
diagnostic tests, and surgical procedures. Using Medicare cost data estimates, the mean cost (Chasan, Maa, Lynch); Center for
incurred during a 2-year period per patient seen in the eye clinic was approximately $1000. Visual and Neurocognitive
Rehabilitation, Atlanta, Georgia
(Delaune); Ophthalmology Section,
CONCLUSIONS AND RELEVANCE Although a teleretinal screening program can be accurate and Atlanta Veterans Affairs Medical
sensitive for multiple visually significant diagnoses, measurable resource burdens should be Center, Decatur, Georgia (Maa,
Lynch).
anticipated to adequately prepare for the associated increase in clinical care.
Corresponding Author: Mary G.
Lynch, MD, Ophthalmology Section
JAMA Ophthalmol. 2014;132(9):1045-1051. doi:10.1001/jamaophthalmol.2014.1051 (112), Atlanta VA Medical Center, 1670
Published online May 29, 2014. Clairmont Rd, Decatur, GA 30033
(mary.lynch4@va.gov).

T
he Veterans Health Administration is the largest inte-
grated health care system in the United States. It serves
more than 5.5 million veterans, including more than 97
million outpatient visits.1 To improve access to health care for
veterans, a diabetic teleretinal screening program was estab-
lished in 2006 to screen veterans for diabetic retinopathy within
community-based primary care clinics. Retinal cameras are
used to capture images, which are remotely interpreted by an
eye care professional in a centralized reading center. Patients
with findings suggestive of ocular disease are referred for an
ophthalmic evaluation in the eye clinic. The diabetic telereti-
nal screening program has improved the provision of preven-
tive screening care to veterans, with approximately 90% of pa-
tients with diabetes mellitus evaluated on a regular basis
nationally.1
The ophthalmic care of patients who are referred by a
diabetic screening program uses resources at medical centers
that include clinic appointments, diagnostic procedures,
spectacles, medications, vision rehabilitation, and surgery.
Consequently, the establishment of a teleretinal imaging ser-
vice significantly increases specialty workload in the affili-
ated eye clinic because, without this service, many veterans
might forgo ophthalmic care. To our knowledge, the effect of
additional referrals on medical center resources has not been

jamaophthalmology.com JAMA Ophthalmology September 2014 Volume 132, Number 9 1045

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 08/25/2019

 Research Original Investigation
evaluated, and this will be important information as the US
Department of Veterans Affairs (VA) contemplates expanding
the teleretinal screening program. In addition, teleophthal-
mology is becoming an important method of health care pro-
vision outside the VA. Data on clinical access and resources
will be crucial for future planning on a national level. There-
fore, this study aimed to evaluate the effect of a community-
based diabetic teleretinal screening program on eye care use
and resources.
Methods
The current study was a retrospective medical record review
of patients who underwent diabetic teleretinal screening in
the community-based clinics of the Atlanta VA Medical Cen-
ter from October 1, 2008, through March 31, 2009, and who
were referred for an ophthalmic examination in the eye
clinic. The data were used to determine the following: (1) the
reasons for referral, (2) the agreement between teleretinal
reads and face-to-face visits, (3) the resource burden per
patient referral for the 2 years after teleretinal imaging, and
(4) possible barriers to patient care. Informed consent was
not obtained because the project was a retrospective medical
record review. All data were collected and stored in a
password-protected file on a secure research-only server
stored in the VA Medical Center.
Inclusion Criteria
This study received institutional review board approval from
the Atlanta VA Medical Center and Emory University and con-
formed to the requirements of the US Health Insurance Por-
tability and Privacy Act. The Veterans Health Information Sys-
tems and Technology Architecture database was used to
identify patients who underwent teleretinal imaging from Oc-
tober 1, 2008, through March 31, 2009, to screen for diabetic
retinopathy. For those patients who had ocular findings that
resulted in a referral to the eye clinic for an ophthalmic exami-
nation, the Computerized Patient Record System (CPRS) was
retrospectively reviewed.
Demographics
Patient age, sex, race, and home zip code were recorded.
Teleretinal imaging notes were reviewed for the referring
diagnoses, which were categorized as nonmacular diabetic
retinopathy, diabetic macular edema, nerve-related disease,
age-related macular degeneration, lens or media opacity, and
other. Patients were included in multiple categories if the
imaging notes contained more than one referring diagnosis.
The CPRS notes were reviewed for patients who were seen in
the eye clinic during the subsequent 2 years. Diagnoses were
grouped in the same manner as the reason for referral.
Only visually significant and/or sight-threatening diagno-
ses were included in the study. For example, dry eye without
corneal changes and dermatochalasis were excluded. Cata-
ract was included only in cases in which it was a referral diag-
nosis or in which the patient subsequently underwent cata-
ract surgery.
1046 JAMA Ophthalmology September 2014 Volume 132, Number 9
Copyright 2014 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 08/25/2019
Teleretinal Screening Program
Accuracy
Teleretinal screening accuracy was determined by the per-
centage of agreement and sensitivity. The percentage of agree-
ment of the teleretinal imaging program was calculated by com-
paring the referral diagnosis to the confirmation diagnosis by
disease category.2 The percentage of agreement was mea-
sured as the number of diagnosis pairs that matched divided
by the total number of diagnosis pairs. Sensitivity was calcu-
lated for each category by dividing the total number of refer-
ral diagnoses confirmed by ophthalmic examination by the
number of diagnoses detected on ophthalmic examination. Pa-
tients were included in multiple categories if appropriate.
Not all patients who were referred for an ophthalmic ex-
amination kept their clinic appointment. For those patients who
were not seen at the Atlanta VA Medical Center during the study
period, the CPRS was searched for the presence of VA notes out-
side the study range or non-VA (community) notes. If such
notes were available, the diagnoses (eg, long-standing age-
related macular degeneration) were collected and added to the
final diagnosis list. Study images were marked as unreadable
if each set was not adequate to determine the presence or ab-
sence of disease. Whether an ocular diagnosis was first de-
tected through the teleretinal screening program was re-
corded on the basis of clinic notes and patient history.
Resource Burden
We recorded the medical center resources that were used dur-
ing the 2-year period after referral from teleretinal screening.
These resources included the total numbers of clinic visits, di-
agnostic procedures, spectacles ordered, medications pre-
scribed, and surgical procedures.
Eye evaluation and management codes were estimated by
allocating the total number of visits in the following manner:
the first visit was considered a new comprehensive eye visit,
the second visit was considered an intermediate established
eye visit, the third visit was considered an established com-
prehensive visit, and subsequent visits were considered in-
termediate established visits.
Resource data were converted to present-day VA mon-
etary cost and relative value units (RVUs) using the standard
2012 Medicare reimbursement and conversion tables and phy-
sician fee schedule search tool (http://www.cms.gov/apps
/physician-fee-schedule/overview.aspx). The mean cost and
RVU workload were then calculated for the number of pa-
tients seen in the eye clinic and the number of patients who
underwent imaging during primary care. Medication and spec-
tacles were not included in the cost calculations.
Barriers to Care
Data were collected from the Veterans Health Information Sys-
tems and Technology Architecture database scheduling menu
to identify possible reasons that a patient would fail to keep
the ophthalmology clinic appointment. The total number of
VA outpatient appointments in the 2 years before teleretinal
screening and the total number of times the patient failed to
keep an appointment during that same period were collected
for all patients referred from a teleretinal screening program.
The ratio of outpatient visit no-shows to total outpatient ap-
jamaophthalmology.com
 Teleretinal Screening Program Original Investigation Research

pointments was then calculated. The shortest driving dis-

Table 1. Patient Demographics and Referral Diagnoses
tance between each patient’s home and the medical center was
calculated using an online trip calculator (Google Maps) and Characteristic Findinga
recorded in miles. Patients screened 1935
Patients referred to eye clinic 465 (24.0)

Teleretinal Imaging Protocol Referred patient age, y

Teleretinal images were obtained using nonmydriatic retinal Mean 64.6
cameras operated by trained licensed practical nurses in pri- Median 64.0
mary care clinics. Most patients had pupillary dilation with a Referred patient race
drop each of phenylephrine hydrochloride, 2.5%, and tropi- White 240 (50.9)
camide, 1%. The screening protocol consisted of 3 wide-angle Black 213 (45.1)
views of the retina and 1 external photograph. The first reti- Other 3 (0.6)
nal photograph was centered on the disc and macula, the sec- Unknown 9 (1.9)
ond along the superotemporal arcade, and the third over the Referred patient sex
nasal retina. Photographs were remotely read by eye care pro- Male 456 (98.1)
fessionals, and a report was generated in the CPRS. Patients Female 9 (1.9)
with an abnormal finding were referred to the local eye clinic Patients seen in eye clinic within 2 years 260 (55.9% of referred;
for further treatment. The eye care professionals followed a 13.4% of screened)
protocol of referral guidelines based on the preferred prac- Ocular notes available (VA or community) 326 (70.1% of referred)
tice patterns of the American Academy of Ophthalmology.3 Referral diagnoses 465
Nonmacular diabetic retinopathy 201 (43.2)
Statistical Analysis Nerve-related disease 143 (30.8)
Statistical analysis was performed using SPSS statistical soft- Lens or media opacity 89 (19.1)
ware (SPSS Inc).4 Most of the information was summarized Age-related macular degeneration 60 (12.9)
descriptively, although some basic exploratory analyses Diabetic macular edema 26 (5.6)
were undertaken to estimate effect size associations among Other 67 (14.4)
variables. Unreadable 45 (9.7% of referred;
2.3% of screened)
Abbreviation: VA, US Department of Veterans Affairs.
a
Data are presented as number (percentage) of patients unless otherwise
Results indicated.
From October 1, 2008, through March 31, 2009, a total of 1935
veterans underwent diabetic teleretinal screening in the pri- Among the 260 patients evaluated in the eye clinic, 42
mary care community-based clinics of the Atlanta VA Medi- (16.2%) had visual acuity of 20/70 or worse in at least 1 eye. Five
cal Center (Table 1). Of those screened, 465 (24.0%) were re- patients (1.9%) were legally blind (visual acuity, <20/200) in
ferred to the eye clinic for an ophthalmic examination. Of those both eyes.
referred, 260 (55.9%) underwent an ophthalmic examination The percentage of agreement for all diagnoses was 90.4%,
within 2 years of the teleretinal screening. Most of the pa- and total sensitivity was 73.6% (Table 2). Images were unread-
tients were male (98.1%); approximately half of the patients able for 45 patients referred from teleretinal screening.
were white and half were black. Ophthalmic notes were avail- Figure 1 shows the mean resource use per patient by pri-
able for an additional 66 patients who were not seen at the VA mary diagnosis category for the 260 patients who had an oph-
during the study period. These notes were reviewed to docu- thalmic examination in the eye clinic during a 2-year period
ment a confirmatory diagnosis but were not included in the after teleretinal imaging. Overall, 109 patients (41.9%) re-
resource allocation data collection. Thus, a confirmatory di- quired only 1 clinic visit, 57 patients (21.9%) required 2 visits,
agnosis was available for 326 (70.1%) of the referred patients. and 94 patients (36.2%) had 3 or more visits. The treatment of
The most common reasons for referral were nonmacular diabetic macular edema had the highest resource use. Pa-
diabetic retinopathy (43.2%), nerve-related disease (30.8%), tients with diabetic macular edema had a mean of 4.96 clinic
lens or media opacity (19.1%), age-related macular degenera- visits, 5.88 diagnostic procedures, and 2.04 surgical proce-
tion (12.9%), and diabetic macular edema (5.6%). Fifty-five pa- dures during the 2-year follow-up period.
tients (16.9%) had 2 or more concurrent problems that put them Diagnostic and surgical procedures were further ana-
at high risk for permanent visual loss (glaucoma, diabetic reti- lyzed to determine the specific rate of use for individual pro-
nopathy, diabetic macular edema, and age-related macular de- cedures as 2 distinct ratios: per patient seen in the eye clinic
generation). and per patient undergoing imaging in primary care (Table 3).
Two categories showed wide racial disparity: 82.2% of pa- Other than refraction, the most common diagnostic proce-
tients with nerve-related disease were African American, and dures were visual field testing and macular optical coherence
91.7% of patients with age-related macular degeneration were tomography. The most common surgical procedures were cata-
white. A visually significant condition was detected for the first ract extraction with intraocular lens insertion, focal laser, pan-
time through teleretinal screening in 142 patients (43.6%). retinal photocoagulation, and intravitreal injection.

jamaophthalmology.com JAMA Ophthalmology September 2014 Volume 132, Number 9 1047

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 08/25/2019

 Research Original Investigation Teleretinal Screening Program

The approximate cost to provide ophthalmic care to this

cohort of patients was estimated through Medicare physi- Discussion
cian fee schedules. The total cost was $251 874.94, with a
mean of $968.75 per patient seen in the clinic and $130.17 per Telehealth technology has been advocated as a method to ex-
patient undergoing imaging in primary care. Diabetic macu- tend clinical care to remote patient communities. Teleretinal
lar edema was the most costly condition to treat (Figure 2). screening has been at the forefront of this initiative. The cur-
The total workload entailed for ophthalmic care was 4190.48 rent study has provided valuable information regarding re-
RVUs, which equaled 16.12 RVUs per patient seen in the source allocation that can be used to prepare for teleretinal ex-
clinic. pansion. Although previous studies5-7 have reported that
Of the 465 patients referred from a teleretinal screening teleretinal screening is generally cost-effective, these studies
program for an eye examination, 205 (44.1%) were not evalu- were based on economic modeling to determine whether a pro-
ated in the clinic. Of the referred patients, 101 (21.7%) did not gram was monetarily responsible. Our study instead col-
show for the scheduled appointment. lected information on actual resources used to predict the fu-
There was a statistically significant difference in the his- ture burden on a health care system. The mean 2-year cost in
torical no-show rate between patients who kept their Medicare reimbursement fees to care for a patient with an ab-
appointment and those who did not (F2,462 = 5.369, P = .005, normal teleretinal screening result who was seen in the eye
I2 = 0.023, 1-way analysis of variance). A patient who histori- clinic was approximately $1000.
cally missed 26% of their scheduled combined outpatient The cost of medication and spectacles was not included
clinic visits had a nearly 57.7% chance of not showing for in the resource burden analyses, which likely led to an under-
the eye clinic visit. There was no statistically significant dif- estimation of the total cost burden. With medication and spec-
ference in the age or the mean driving distance for patients tacles, the actual cost varies among VA medical centers de-
who kept their eye clinic appointment vs patients who did pending on local policy and contracts. Some medication is
not show. considered over the counter in the community and not cov-
ered by insurance but might be covered by the VA pharmacy.
Table 2. Accuracy of Teleretinal Screening in Detecting Similarly, spectacles often are not covered by insurance, but
Diagnosis Categories a basic style of spectacles is covered by the VA. Thus, the raw
All Cases With Confirmatory Diagnosis numbers for medication and spectacles were included to al-
(N = 326) low for some predictive estimates of these important resources.
Diagnosis Agreement, % Sensitivity, % Other costs inherent in and paid for by the VA health care
Nerve-related disease 91.1 81.2 system, such as parking, security, and information technol-
Nonmacular diabetic retinopathy 89.9 88.4 ogy, are shared by the entire enterprise and could not be in-
Diabetic macular edema 94.5 56.0 cluded in this analysis. In addition, the costs incurred by imple-
Age-related macular 95.4 81.6 menting a teleretinal program (eg, cameras, computers,
degeneration
imagers, and readers) were not considered in this analysis of
Lens or media opacity 91.7 75.3
resources needed after a screening visit. Consequently, the total
Other 79.8 36.6
estimate of care determined by the Medicare physician fee
Total 90.4 73.6
schedules does not reflect the total cost to the VA.

Figure 1. Eye Care Resource Use by Diagnostic Category for 2 Years After Teleretinal Screening

7.00

6.00

5.00
Mean Use

4.00

3.00

2.00

1.00

0
Nerve-Related Age-Related Nonmacular Diabetic Lens or Other
Disease Macular Diabetic Macular Edema Media Opacity
Degeneration Retinopathy The bar graph shows the mean
resource use per patient by primary
Cumulative No. of visits 2.11 2.03 2.79 4.96 3.52 2.89
diagnosis category for the 260
Total No. of spectacles 1.20 0.78 1.10 0.96 1.22 1.07
Total No. of prescriptions 0.39 0.57 0.46 0.84 1.18 0.69 patients who underwent an
Total No. of diagnostic procedures 3.50 2.22 3.01 5.88 3.29 3.46 ophthalmic examination in the eye
Total No. of surgical procedures 0.11 0.24 0.59 2.04 0.74 0.47 clinic during a 2-year period after
teleretinal imaging.

1048 JAMA Ophthalmology September 2014 Volume 132, Number 9 jamaophthalmology.com

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 08/25/2019

 Teleretinal Screening Program Original Investigation Research

Table 3. Diagnostic and Surgical Use Ratios

Total No. of Rate per Imaged Patient Rate per Patient Seen in Clinic
Procedure Patients (n = 1935) (n = 260)
Diagnostic procedure
A-scan 1 0.000517 0.003846
B-scan 22 0.011370 0.084615
Pachymetry 65 0.033592 0.250000
IOL calculation 24 0.012403 0.092308
Refraction 410 0.211886 1.576923
Gonioscopy 74 0.038243 0.284615
Topography 4 0.002067 0.015385
Bandage contact lens 1 0.000517 0.003846
GVF 10 0.005168 0.038462
HVF 118 0.060982 0.453846
OCT nerve 21 0.010853 0.080769
OCT macula 91 0.047028 0.350000
Fluorescein angiography 43 0.022222 0.165385
Fundus photography 6 0.003101 0.023077
External photography 1 0.000517 0.003846
Contact lens 5 0.002584 0.019231
Surgical procedure
Suture removal 11 0.005685 0.042308
Laser iridotomy 1 0.000517 0.003846
YAG capsulotomy 4 0.002067 0.015385
Cataract extraction 31 0.016021 0.119231
Fluid-gas exchange 1 0.000517 0.003846
Vitreous injection 8 0.004134 0.030769
Vitrectomy 3 0.001550 0.011538
Focal endolaser 1 0.000517 0.003846
Membrane strip 1 0.000517 0.003846
Scleral buckle with 1 0.000517 0.003846
cryotherapy
External laser for RD 1 0.000517 0.003846
Focal laser 27 0.013953 0.103846
Abbreviations: GVF, Goldmann visual
PRP 24 0.012403 0.092308
field; HVF, Humphrey visual field;
Tenon injection 1 0.000517 0.003846 IOL, intraocular lens; OCT, optical
Excision of eyelid lesion 1 0.000517 0.003846 coherence tomography;
PRP, panretinal photocoagulation;
Entropion repair 1 0.000517 0.003846
RD, retinal detachment.

Figure 2. Mean 2-Year Cost per Patient Seen in the Eye Clinic

3000.00
$2673.36

2500.00
Mean Cost per Patient, $

2000.00

1500.00
$1184.06
$1038.33 $1027.90
1000.00
$602.41 The bar graph shows the mean
$469.42 approximate cost to provide
500.00
ophthalmic care to patients referred
from diabetic teleretinal screening as
0
Nerve-Related Age-Related Nonmacular Diabetic Lens or Other estimated through Medicare
Disease Macular Diabetic Macular Edema Media Opacity physician fee schedules. Diabetic
Degeneration Retinopathy macular edema was the most costly
condition to treat.

jamaophthalmology.com JAMA Ophthalmology September 2014 Volume 132, Number 9 1049

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 08/25/2019

 Research Original Investigation Teleretinal Screening Program

Use ratios generated by this study can be used to predict
the number of appointment slots, diagnostic tests, and surgi-
cal procedures that will be needed to care for patients who are
referred for an ophthalmic examination. For example, plac-
ing a teleretinal camera in a primary care clinic that serves a
population that includes 5000 diabetic patients may gener-
ate approximately 1200 referrals to an eye clinic. Assuming that
all patients are evaluated with an ophthalmic examination,
these referrals might require approximately 544 visual fields,
516 optical coherence tomography images, and 143 cataract ex-
tractions during the next 2 years. A sizeable appointment no-
show rate for these referrals might adjust these estimates down.
The ratio data obtained from this study may be further ex-
panded using existing actuarial data for populations not
screened currently. Combining our data on resource burden
with population size and diabetes prevalence should allow de-
cision makers to appropriately expand eye clinic appoint-
ments, hire additional personnel, and purchase equipment in
anticipation of an increasing patient population.
The findings of the present study add further support to
the effectiveness of teleretinal screening for diabetic eye
disease.8-11 Although teleretinal screening has yet to be vali-
dated for macula and optic nerve disease, diabetic patients with
nondiabetic findings currently are being referred for a clinical
examination, and teleretinal screening may be a useful method
to detect other ocular conditions. The value of diabetic telereti-
nal imaging conducted in a primary care setting is high. Of the
326 patients for whom a note was available, 142 patients (43.6%)
had an ocular condition detected for the first time through
teleretinal screening. Fifty-five of 326 patients (16.9%) had 2 or
more concurrent problems that put them at high risk for vi-
sual loss (glaucoma, diabetic retinopathy, diabetic macular
edema, and age-related macular degeneration). Because many
ocular conditions are both progressive and treatable, this form
of screening can be effective in identifying patients with ocu-
lar disease in order to initiate sight-saving care.
Unfortunately, a high percentage of patients (44.1%) re-
ferred for ophthalmic examinations were not evaluated in the
eye clinic. Of the 465 patients who were referred for evalua-
tion, 101 (21.7%) failed to show for the scheduled appoint-
ment. Failing to show for the appointment puts the patient at
risk for delay in care and wastes valuable medical center re-
sources. The current study found that age and travel distance
were not significant factors that contributed to the no-show
rate. Rather, the patient’s historical no-show rate was the best
predictor of whether the patient would keep the eye clinic
appointment.
Predictive modeling for eye care will be an important strat-
egy in the management of health care disparity. A major con-
tributor to increased eye care workload in any health care sys-
tem is the care of a patient with diabetic retinopathy, which is
one of the leading causes of preventable blindness.12 The vet-
eran population has a diabetes prevalence of approximately
20%.13 Early detection and treatment of diabetic eye disease
lead to a reduction in moderate to severe visual loss14 and may
save the federal government hundreds of millions of dollars.7
Our study suggests that teleretinal imaging may also be used
to detect other sight-threatening conditions, and the ophthal-
mic community should be prepared for increased workload and
resource use.
The study had a number of limitations. First, all patients
were from the Southeastern United States, an area with a high
prevalence of diabetes. Second, patients were elderly, with a
mean age of 65 years, and male, typical of the VA patient popu-
lation. It is likely that a younger cohort of patients with dia-
betes would have a lower prevalence of ocular findings. Con-
sequently, the referral rate from a teleretinal imaging program
that involved younger and healthier patients might be lower.
However, once the referral rate for that population was estab-
lished, the resource use rates for referred patients would likely
be equivalent to those measured in this study. Third, the study
was conducted at a time when the eye clinic at the VA was tran-
sitioning from fluorescein angiography to optical coherence
tomography as the primary means of following up patients with
suspected macular edema. In addition, the surgical care of
macular disease was transitioning from primarily laser based
to include more intravitreal injections. Consequently, the cur-
rent rate of optical coherence tomography use and intravit-
real injections would likely be higher than what was mea-
sured by the study.
Finally, the office visit burden was estimated based on the
number of eye clinic visits a patient had during a 2-year pe-
riod rather than by using specific evaluation and manage-
ment codes. During the time of teleretinal screening and the
subsequent clinical care, all visit codes in the VA system were
manually logged and entered into the code capture system by
clerks. The electronic health record for eye care with auto-
matic code capture that is now being implemented at the VA
was not in use. Thus, at the time of this study, the evaluation
and management coding data were not as reliable at the VA as
they are now. Visit burden was estimated based on the num-
ber of visits for each patient.
Conclusions
Diabetic teleretinal imaging is an effective method to screen
patients for sight-threatening conditions. In addition, there are
measureable resource burdens that can be anticipated to ad-
equately prepare for the associated increase in clinical care.

ARTICLE INFORMATION
Submitted for Publication: August 2, 2013; final
revision received September 5, 2013; accepted
January 31, 2014.
Published Online: May 29, 2014.
doi:10.1001/jamaophthalmol.2014.1051.
Author Contributions: Drs Lynch and Chasan had
full access to all the data in the study and take
responsibility for the integrity of the data and the
accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data:
Chasan, Delaune, Lynch.
Drafting of the manuscript: Chasan, Lynch.
Critical revision of the manuscript for important
intellectual content: All authors.
Statistical analysis: Chasan, Delaune.
Administrative, technical, or material support:
Chasan, Lynch.
Study supervision: Maa, Lynch.
Conflict of Interest Disclosures: None reported.

1050 JAMA Ophthalmology September 2014 Volume 132, Number 9 jamaophthalmology.com

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 08/25/2019

 Teleretinal Screening Program Original Investigation Research

Disclaimer: The views expressed here are those of
the authors and do not necessarily reflect the
position or policy of the US Department of Veterans
Affairs or the US government.
Additional Contributions: Paul B. Greenberg, MD,
Department of Surgery (Ophthalmology), Alpert
Medical School of Brown University, and
Providence Veterans Affairs Medical Center,
Providence, Rhode Island, provided editorial
assistance.
REFERENCES
1. Veterans Health Administration. VHA Support
Service Center [database online]. Austin, TX:
Veterans Health Administration. http://vssc.med
.va.gov. Accessed December 12, 2012.
2. Hunt RJ. Percent agreement, Pearson’s
correlation, and kappa as measures of inter-examiner
reliability. J Dent Res. 1986;65(2):128-130.
3. American Academy of Ophthalmology Retina
Panel. Preferred Practice Pattern® Guidelines:
Diabetic Retinopathy. San Francisco, CA: American
Academy of Ophthalmology; 2008.
4. IBM Corp. IBM SPSS Statistics for Windows,
Version 21.0. Armonk, NY: IBM Corp; 2012.
5. Jones S, Edwards RT. Diabetic retinopathy
screening: a systematic review of the economic
evidence. Diabet Med. 2010;27(3):249-256.
6. Whited JD, Datta SK, Aiello LM, et al. A modeled
economic analysis of a digital tele-ophthalmology
system as used by three federal health care agencies
for detecting proliferative diabetic retinopathy.
Telemed J E Health. 2005;11(6):641-651.
7. Javitt JC, Aiello LP, Chiang Y, Ferris FL III, Canner
JK, Greenfield S. Preventive eye care in people with
diabetes is cost-saving to the federal government:
implications for health-care reform. Diabetes Care.
1994;17(8):909-917.
8. Cavallerano AA, Cavallerano JD, Katalinic P,
Tolson AM, Aiello LP, Aiello LM; Joslin Vision
Network Clinical Team. Use of Joslin Vision Network
digital-video nonmydriatic retinal imaging to assess
diabetic retinopathy in a clinical program. Retina.
2003;23(2):215-223.
9. Gómez-Ulla F, Fernandez MI, Gonzalez F, et al.
Digital retinal images and teleophthalmology for
detecting and grading diabetic retinopathy.
Diabetes Care. 2002;25(8):1384-1389.
10. Kerr D, Cavan DA, Jennings B, Dunnington C,
Gold D, Crick M. Beyond retinal screening: digital
imaging in the assessment and follow-up of
patients with diabetic retinopathy. Diabet Med.
1998;15(10):878-882.
11. Bragge P, Gruen RL, Chau M, Forbes A, Taylor
HR. Screening for presence or absence of diabetic
retinopathy: a meta-analysis. Arch Ophthalmol.
2011;129(4):435-444.
12. Aiello LP, Gardner TW, King GL, et al. Diabetic
retinopathy. Diabetes Care. 1998;21(1):143-156.
13. Orcutt J, Avakian A, Koepsell TD, Maynard C.
Eye disease in veterans with diabetes. Diabetes Care.
2004;27(suppl 2):B50-B53.
14. Early Treatment Diabetic Retinopathy Study
Research Group. Early photocoagulation for
diabetic retinopathy. ETDRS report number 9.
Ophthalmology. 1991;98(5)(suppl):766-785.

OPHTHALMIC IMAGES

Acquired Trichomegaly
Trichomegaly Secondary to Erlotinib
Carlos A. Medina Mendez, MD; Patrick C. Ma, MD; Arun D. Singh, MD

Front (A) and lateral (B) external photographs of a woman in her early 60s who
had undergone treatment with erlotinib for more than 2 and a half years. She
developed bilateral trichomegaly, which required constant trimming.

jamaophthalmology.com JAMA Ophthalmology September 2014 Volume 132, Number 9 1051

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 08/25/2019