Ophthalmic Surgery, Lasers and Imaging Retina

Clinical Science 

Retina Specialists Treating Diabetic Macular Edema Recommend Different Approaches for Patients Than They Would Choose for Themselves

Charles G. Miller, PhD; Greg Budoff, BS; Karen W. Jeng-Miller, MD, MPH; Howard F. Fine, MD, MHSc; Daniel B. Roth, MD; Jonathan L. Prenner, MD

Abstract

BACKGROUND AND OBJECTIVE:

Prior investigation shows retina specialists may select different treatment for age-related macular degeneration for themselves than for a hypothetical patient. The authors sought to investigate whether a similar bias exists for treatment decisions by retina specialists with regard to diabetic macular edema (DME).

PATIENTS AND METHODS:

Two surveys asked retina specialists to select treatment for hypothetical patients with DME or for themselves. In Survey 2, a distinction was drawn between a visual acuity (VA) of 20/40 or better and 20/50 or worse.

RESULTS:

In Survey 1, 54% to 61% of respondents selected bevacizumab (Avastin; Genentech, South San Francisco, CA) for patients and 36% to 40% selected the drug for themselves (P < .0004). It was found that 14% to 17% selected aflibercept (Eylea; Regeneron, Tarrytown, NY) for patients versus 31% to 38% who selected it for themselves (P < .0001). For a VA of 20/40 or better, 42% to 50% selected bevacizumab for their patients versus 32% to 39% (P < .0005) for themselves, and 20% to 23% selected aflibercept for patients versus 39% to 48% (P < .0007) for themselves. For a VA of 20/50 or worse, 24% to 28% chose bevacizumab for patients versus 17% to 20% for themselves (P value was not significant), and 59% to 66% selected aflibercept for their patients versus 66% to 78% for themselves (P < .05).

CONCLUSION:

Physicians recommend different treatment for their patients than for themselves, though not for a VA of 20/50 or worse, where data support the use of aflibercept over bevacizumab.

[Ophthalmic Surg Lasers Imaging Retina. 2016;47:544–554.]

Abstract

BACKGROUND AND OBJECTIVE:

Prior investigation shows retina specialists may select different treatment for age-related macular degeneration for themselves than for a hypothetical patient. The authors sought to investigate whether a similar bias exists for treatment decisions by retina specialists with regard to diabetic macular edema (DME).

PATIENTS AND METHODS:

Two surveys asked retina specialists to select treatment for hypothetical patients with DME or for themselves. In Survey 2, a distinction was drawn between a visual acuity (VA) of 20/40 or better and 20/50 or worse.

RESULTS:

In Survey 1, 54% to 61% of respondents selected bevacizumab (Avastin; Genentech, South San Francisco, CA) for patients and 36% to 40% selected the drug for themselves (P < .0004). It was found that 14% to 17% selected aflibercept (Eylea; Regeneron, Tarrytown, NY) for patients versus 31% to 38% who selected it for themselves (P < .0001). For a VA of 20/40 or better, 42% to 50% selected bevacizumab for their patients versus 32% to 39% (P < .0005) for themselves, and 20% to 23% selected aflibercept for patients versus 39% to 48% (P < .0007) for themselves. For a VA of 20/50 or worse, 24% to 28% chose bevacizumab for patients versus 17% to 20% for themselves (P value was not significant), and 59% to 66% selected aflibercept for their patients versus 66% to 78% for themselves (P < .05).

CONCLUSION:

Physicians recommend different treatment for their patients than for themselves, though not for a VA of 20/50 or worse, where data support the use of aflibercept over bevacizumab.

[Ophthalmic Surg Lasers Imaging Retina. 2016;47:544–554.]

Introduction

It is estimated that 29.1 million Americans are currently living with diabetes mellitus (DM).1 Nearly a third of patients with DM suffer from diabetic retinopathy, making it a leading cause of visual loss in the United States.2 Diabetic macular edema (DME), a consequence of diabetic retinopathy that affects central vision, is estimated to affect 750,000 patients in the U.S.3 In the face of such significant disease burden, it is instructive to understand how clinicians make treatment decisions in the management of DME. The intravitreal injection of anti-vascular endothelial growth factor (VEGF) biologic agents has supplanted laser photocoagulation as the standard of care for DME.4 Three anti-VEGF biologics are currently available for the treatment of DME (aflibercept [Eylea; Regeneron, Tarrytown, NY], bevacizumab [Avastin; Genentech, South San Francisco, CA], and ranibizumab [Lucentis; Genentech, South San Francisco, CA]), and all have been shown to be superior to laser photocoagulation in preserving vision.5 Utilization rates of these therapies for the treatment of DME by retina specialists are of particular interest.

Prior studies have shown that physician recommendations, which greatly influence patients' treatment decisions, are impacted by cognitive biases.6 In fact, retina specialists have been shown to select different treatments for age-related macular degeneration (AMD) for themselves than they would for a hypothetical patient.7 We sought to investigate whether a similar bias exists for treatment decisions by retina specialists with regard to DME. Recent data demonstrate that aflibercept is more effective than other anti-VEGF therapies at improving vision in patients with DME who have a visual acuity (VA) of 20/50 or worse.8 We also sought to determine whether such data have impacted retina specialists' decision-making in treating themselves or a hypothetical patient.

An initial survey (Survey 1) was sent to assess retina specialists' treatment choices for hypothetical patients versus themselves across 11 different case descriptions, irrespective of VA. A subsequent survey (Survey 2) was performed after the publication of The Diabetic Retinopathy Clinical Research Network's findings that demonstrated aflibercept superiority in eyes with 20/50 or worse vision. In this second survey, clinicians were asked to select a therapeutic approach for case descriptions with a VA of 20/40 or better and with a VA of 20/50 or worse. Informed by the existing literature, we hypothesized that respondents would in fact select different therapy for a hypothetical patient than for themselves. We expected for these differences to be diminished in the context of DME patients with a VA of 20/50 or worse, where the most recent data are clear regarding the superiority of aflibercept over other anti-VEGF therapies.

Patients and Methods

Two samples of retina specialists selected from the American Society of Retina Specialists membership database were surveyed as to their treatment recommendations for DME. The survey groups were evenly balanced by location to account for potential regional differences in prescribing habits. We provided each group with a clinical scenario and collected subsequent responses through an electronic survey managed by Survey Monkey software. One group was asked to select therapy for a hypothetical patient and the other to imagine that they themselves were being treated for DME (Survey 1; Table 1). The two questions were: “Patient”: You have a patient with diabetic macular edema that is not amenable to laser photocoagulation. What therapeutic treatment would you choose for each of the following patients? “Self”: Imagine you are a patient with diabetic macular edema that is not amenable to laser photocoagulation. What therapeutic treatment would you choose if you were each of the following patients?


Respondent Characteristics

Table 1:

Respondent Characteristics

Each group selected among seven different treatment options for 11 different patient descriptions. Treatment options included: Eylea/aflibercept; Lucentis/ranibizumab; Avastin/bevacizumab; Kenalog/triamcinolone acetonide (Bristol-Myers Squibb, New York, NY); Triesence/triamcinolone acetonide (Alcon, Fort Worth, TX); Ozurdex/dexamethasone intravitreal implant (Allergan, Irvine, CA); and Iluvien/fluocinolone (Alimera Sciences, Alpharetta, GA). Patient descriptions are outlined in Table 2.


Bevacizumab for “Self” Versus “Patient” in Survey 1

Table 2:

Bevacizumab for “Self” Versus “Patient” in Survey 1

A subsequent survey (Survey 2; Table 1) was performed in which two additional groups of retina specialists, selected in an identical fashion, were asked to choose from among the same treatment options for the same patient descriptions, but each group was asked two sets of questions that specified visual acuity. The questions were as follows:

“Patient”: 1. You have a patient with diabetic macular edema that is not amenable to laser photocoagulation. What therapeutic treatment would you choose for each of the following patients if they had a visual acuity of 20/40 or better?2. You have a patient with diabetic macular edema that is not amenable to laser photocoagulation. What therapeutic treatment would you choose for each of the following patients if they had a visual acuity of 20/50 or worse?

“Self”: 1. Imagine you are a patient with diabetic macular edema that is not amenable to laser photocoagulation. What therapeutic treatment would you choose if you were each of the following patients and had a visual acuity of 20/40 or better?2. Imagine you are a patient with diabetic macular edema that is not amenable to laser photocoagulation. What therapeutic treatment would you choose if you were each of the following patients and had a visual acuity of 20/50 or worse?

Covariates and Data Analysis

In both surveys, the respondents provided answers concerning their practice setting, location of practice, years of practice, and gender. Chi-square tests were performed to compare choices between the “self” group versus the “patient” group. Multivariable logistic regression analysis was performed with choice of treatment as the dependent variable and “self” versus “patient” as the primary independent variable. Additional multivariable logistic regression analyses were performed with choice of treatment as the dependent variable and each of the four demographic variables as primary independent variables. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).

Results

An initial survey (Survey 1) was sent to 2,419 retina specialists and 410 responses were returned, for a response rate of 17%. The first group of 199 respondents selected therapy for a hypothetical patient with DME. These data were compared to answers from the second group of 211 respondents, who were asked to select a therapy to treat themselves in the event that they carried a diagnosis of DME. Despite a relatively low response rate, respondent characteristics of the “patient” and “self” groups were similar (Table 1), thus reducing the potential for sampling bias. Treatment options included several intravitreal anti-VEGF therapies: aflibercept, bevacizumab, and ranibizumab.

Results from this survey demonstrate that retina specialists would select different therapies for a hypothetical patient than they would for themselves (Figure 1). Anti-VEGF medications were the overwhelming choice of therapy for all case descriptions, except those of “anti-VEGF nonresponders,” for whom intravitreal steroids of various preparations were favored. Respondents chose bevacizumab overwhelmingly to treat hypothetical patients with DME, except in the case of anti-VEGF nonresponders (Figure 1A). When compared to the “self” group, respondents in the “patient” group had a statistically significant (P < .0004) preference for bevacizumab when treating patients of all descriptions except for anti-VEGF nonresponders (Table 2). Meanwhile, respondents in the “self” group were more likely to select aflibercept (Figure 1B) than those in the “patient” group (P < .0001) (Table 3). Respondents from the Northeast were also more likely to select bevacizumab (P < .002) than respondents from other regions (South, Midwest, or West Coast) for all patient descriptions except for anti-VEGF non-responders (Table 4). Geography also had a statistically significant effect on use of aflibercept and ranibizumab in four of the 11 case descriptions (data not shown). No other respondent characteristics were found to influence treatment selection.


Retina specialists select different therapy for themselves than for a hypothetical patient when treating diabetic macular edema. We surveyed 2,419 retina specialists, of whom 410 responded as to treatment recommendations for 11 different patient descriptions (X-axis) and percentage selecting each therapy are shown (Y-axis). Results are shown for respondents asked to select therapy for a hypothetical patient (A) and for themselves (B). DME = diabetic macular edema; y/o = years old; BDR = background diabetic retinopathy; PDR = proliferative diabetic retinopathy; VEGF = vascular endothelial grownth factor

Figure 1.

Retina specialists select different therapy for themselves than for a hypothetical patient when treating diabetic macular edema. We surveyed 2,419 retina specialists, of whom 410 responded as to treatment recommendations for 11 different patient descriptions (X-axis) and percentage selecting each therapy are shown (Y-axis). Results are shown for respondents asked to select therapy for a hypothetical patient (A) and for themselves (B). DME = diabetic macular edema; y/o = years old; BDR = background diabetic retinopathy; PDR = proliferative diabetic retinopathy; VEGF = vascular endothelial grownth factor


Aflibercept for “Self” Versus “Patient” in Survey 1

Table 3:

Aflibercept for “Self” Versus “Patient” in Survey 1


Geographic Distribution of Bevacizumab Use in Survey 1

Table 4:

Geographic Distribution of Bevacizumab Use in Survey 1

A second survey (Survey 2) was sent to 2,669 retina specialists and 216 responses were received, yielding a response rate of 8.1%. In contrast to Survey 1, respondents were asked two sets of questions: one regarding patients with DME and a visual acuity (VA) of 20/40 or better (Question 1) and another set with a VA of 20/50 or worse (Question 2). Again, one group (107 respondents) was asked to select therapy for a hypothetical patient with DME and another group was asked to select therapy as if they were treating themselves (109 respondents). Respondent characteristics between the “patient” and “self” groups were again compared (Table 1).

Results from Survey 2 show that VA impacts respondents' choice of therapy (Figure 2). In contrast to Survey 1, there was no statistically significant difference between respondents' use of bevacizumab between the “patient” and “self” groups for all but one patient description (“phakic, 35 years old, and has moderate glaucoma and BDR”) with a VA of 20/40 or better. For a VA of 20/50 or worse, there was no statistically significant difference in the use of bevacizumab (Table 5) for any of the patient descriptions. The tendency for retina specialists from the Northeast to select bevacizumab, observed in Survey 1, persisted in Survey 2 for patients with a VA of 20/40 or better in all cases except for anti-VEGF nonresponders (P < .05) (Table 6). There was no such influence of geography on treatment decisions for a VA of 20/50 or worse.


Treatment recommendations for diabetic macular edema by retina specialists differ for patients with visual acuity (VA) of 20/40 or better versus those with VA of 20/50 or worse. Of the 2,669 retina specialists surveyed, 216 responded as to treatment recommendations for 11 different patient descriptions (X-axis) with VA of 20/40 or better and VA of 20/50 or worse. The percentage of respondents selecting each therapy are shown (Y-axis). Results are shown for respondents asked to select therapy for a hypothetical patient (A) and for themselves (B). DME = diabetic macular edema; y/o = years old; BDR = background diabetic retinopathy; PDR = proliferative diabetic retinopathy; VEGF = vascular endothelial grownth factor

Figure 2.

Treatment recommendations for diabetic macular edema by retina specialists differ for patients with visual acuity (VA) of 20/40 or better versus those with VA of 20/50 or worse. Of the 2,669 retina specialists surveyed, 216 responded as to treatment recommendations for 11 different patient descriptions (X-axis) with VA of 20/40 or better and VA of 20/50 or worse. The percentage of respondents selecting each therapy are shown (Y-axis). Results are shown for respondents asked to select therapy for a hypothetical patient (A) and for themselves (B). DME = diabetic macular edema; y/o = years old; BDR = background diabetic retinopathy; PDR = proliferative diabetic retinopathy; VEGF = vascular endothelial grownth factor


Bevacizumab for “Self” Versus “Patient” in Survey 2

Table 5:

Bevacizumab for “Self” Versus “Patient” in Survey 2


Geographic Distribution of Bevacizumab Use in Survey 2

Table 6:

Geographic Distribution of Bevacizumab Use in Survey 2

The use of aflibercept, on the other hand, followed a pattern similar to that of Survey 1 when respondents were asked to consider therapy for a hypothetical patient or themselves with a VA of 20/40 or better. There was a statistically significant (P < .0007) preference for aflibercept in the “self” group as compared to the “patient” group across all case descriptions with a VA of 20/40 or better, except for anti-VEGF nonresponders (Table 7). This preference for aflibercept in the “self” group persisted (P < .05) for all but two patient descriptions with a VA of 20/50 or worse (Table 7). There was a statistically significant (P < .0007) preference for ranibizumab for “patient” versus “self” in Survey 2, with a VA of 20/40 or better for all patient descriptions except anti-VEGF nonresponders (Table 8). This difference was also seen (P < .05) for all but three patient descriptions for a VA of 20/50 or worse (Table 8). This was a departure from the results seen in Survey 1, where there was no statistically significant difference in the use of ranibizumab between “patient” and “self” groups.


Aflibercept for “Self” Versus “Patient” in Survey 2

Table 7:

Aflibercept for “Self” Versus “Patient” in Survey 2


Ranibizumab for “Self” Versus “Patient” in Survey 2

Table 8:

Ranibizumab for “Self” Versus “Patient” in Survey 2

Discussion

The increasing prevalence of DM and its complications, including retinopathy and resultant DME, necessitate a more complete understanding of the factors that influence practitioners' choice of management. These data demonstrate that retina specialists select different therapeutic strategies for hypothetical patients than for themselves. This was an expected result in light of previous studies that demonstrated a similar phenomenon with selection of anti-VEGF therapy for the treatment of exudative AMD.7 The results of Survey 2, however, are of particular interest. In the modern age of evidence-based medicine, randomized, controlled trials (RCTs) are viewed as the primary effector of change to patterns of practice. In the field of cardiology, for example, the impact of major drug trials on clinical practice has been verified.9 Studies evaluating the efficacy of such trials in instigating such changes are lacking in the field of ophthalmology. More than 30 years ago, the U.S. Office of Technology Assessment undertook a concerted effort to evaluate the impact of RCTs on clinical decision-making in medicine more broadly.10 This background paper singled out the National Eye Institute for effective study design.11 In the absence of such a far-reaching critical appraisal in the current era, it is important to judge the impact of such trials once the findings have been disseminated.

The results of Survey 2 suggest that the data published by The Diabetic Retinopathy Clinical Research Network in March 2015 regarding the relative efficacy of aflibercept, bevacizumab, and ranibizumab in treating DME are in line with retina specialists' responses to our survey questions. We found that retina specialists were most likely to recommend aflibercept for the treatment of DME with a VA of 20/50 or worse whether asked about treatment for a hypothetical patient or themselves (Figure 2). This trend may in part be attributable to the aforementioned study's conclusion that “at worse initial levels of vision, aflibercept had a clinically meaningful advantage.”8 Since Survey 1 did not include questions regarding specific VA, the effect of this conclusion cannot be measured directly. Interestingly, there was a statistically significant difference between the use of bevacizumab for treating “self” vs. “patient” in Survey 2 for a VA of 20/40 or better, but no such difference was seen for a VA of 20/50 or worse. Survey 1 documented a similar bias against the use of bevacizumab for the “self” group. Thus, any factors that drove retina specialists to select alternatives to bevacizumab when treating themselves in Survey 1 likely persisted for Survey 2 with regard to patients with a VA of 20/40 or better. In contrast, bevacizumab was an unpopular choice for a VA of 20/50 or worse whether for “self” or “patient,” likely due to the data-supported preference for aflibercept cited above.

Previous studies of patterns of use for anti-VEGF therapy by retina specialists have demonstrated that the geographical area in which physicians practice impacts their choice of therapy in treating AMD.7 Interestingly, we show a similar phenomenon with regard to choice of anti-VEGF therapy when treating DME. Practitioners in the Northeast were found to be more likely to select bevacizumab in Survey 1 (P < .0002) and in Survey 2 for a VA of 20/40 or better (P < .05). This trend runs counter to that observed by Jeng et al., where respondents in the South, Midwest, and West Coast were all more likely to select bevacizumab for treatment of AMD than physicians in the Northeast. Possible explanations for regional differences in treatment patterns for DME include variations in the incidence of diabetes and thus its resultant complications across the U.S. The South has been noted to have markedly higher rates of diabetes as compared to the rest of the U.S.12 Again, we see that whatever factors may be at play in driving the geographical discrepancy in treatment patters are no longer present when cases with a VA of 20/50 or worse are considered. Here, the data are clearer and treatment patterns are appropriately more consistent across different regions.

Overall, this study confirms that retina specialists select different therapy for a hypothetical patient than for themselves when treating DME with all other factors being equal. We also document that these discrepancies are less pronounced for patients with a VA of 20/50 or worse. A greater sensitivity to cost of treatment when considering choice of therapy on behalf of patients, rather than for themselves, may explain some of the discrepancies in treatment selection seen in Survey 1, as was suggested by Jeng et al. with regard to therapy for AMD.7 In addition to being a less expensive treatment option than other available anti-VEGF therapies, bevacizumab has been shown to be more cost-effective in combination with photocoagulation than ranibizumab for the management of DME.13 Since aflibercept is even more costly than ranibizumab, these observations can be extended to include it, as well. This discrepancy has provoked a debate regarding the circumstances under which the benefit to the patient justifies the increased cost and, more importantly, how to make the costlier option more widely available to patients who stand to benefit.14,15 Further studies will be necessary to help answer these questions, and it will be imperative to continue to assess the impact of such studies on treatment patterns.

References

  1. Centers for Disease Control and Prevention. National diabetes statistics report: estimates of diabetes and its burden in the United States, 2014. Available at http://www.cdc.gov/diabetes/pubs/statsreport14/national-diabetes-report-web.pdf. Accessed September 8, 2015.
  2. Zhang X, Saaddine JB, Chou CF, et al. Prevalence of diabetic retinopathy in the United States, 2005–2008. JAMA. 2010;304(6):649–656. doi:10.1001/jama.2010.1111 [CrossRef]
  3. Varma R, Bressler NM, Doan QV, et al. Prevalence of and risk factors for diabetic macular edema in the United States. JAMA Ophthalmol. 2014;132(11):1334–1340. doi:10.1001/jamaophthalmol.2014.2854 [CrossRef]
  4. Agarwal A, Afridi R, Hassan M, et al. Novel therapies in development for diabetic macular edema. Curr Diab Rep. 2015;15(10):652. doi:10.1007/s11892-015-0652-z [CrossRef]
  5. Boyer DS, Hopkins JJ, Sorof J, Ehrlich JS. Anti-vascular endothelial growth factor therapy for diabetic macular edema. Ther Adv Endocrinol Metab. 2013;4(6):151–169. doi:10.1177/2042018813512360 [CrossRef]
  6. Ubel PA, Angott AM, Zikmund-Fisher BJ. Physicians recommend different treatments for patients than they would choose for themselves. Arch Intern Med. 2011;171(7):630–634. doi:10.1001/archinternmed.2011.91 [CrossRef]
  7. Jeng KW, Wilgucki J, Halperin S, et al. Retina specialists treating age-related macular degeneration recommend different approaches for patients than they would choose for themselves. Retina. 2014;34(9):1796–1801. doi:10.1097/IAE.0000000000000182 [CrossRef]
  8. Wells JA, Glassman AR, Diabetic Retinopathy Clinical Research Network et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema. N Engl J Med. 2015;372(13):1193–1203. doi:10.1056/NEJMoa1414264 [CrossRef]
  9. Lamas GA, Pfeffer MA, Hamm P, Wertheimer J, Rouleau JL, Braunwald E. Do the results of randomized clinical trials of cardiovascular drugs influence medical practice? The SAVE Investigators. N Engl J Med. 1992;327(4):241–247. doi:10.1056/NEJM199207233270405 [CrossRef]
  10. Office of Technology Assessment. The Impact of Randomized Clinical Trials on Health Policy and Medical Practice: Background Paper. Vol OTA-BP-H-22. Washington, DC;1983:1–103.
  11. Goldberg IA. The impact of randomized clinical trials on health policy and medical practice. Arch Ophthalmol. 1985;103(3):332. doi:10.1001/archopht.1985.01050030028014 [CrossRef]
  12. Centers for Disease Control and Prevention. Diabetes Report Card 2012. Available at http://www.cdc.gov/diabetes/pubs/pdf/diabetesre-portcard.pdf. Accessed September 8, 2015.
  13. Stein JD, Newman-Casey PA, Kim DD, Nwanyanwu KH, Johnson MW, Hutton DW. Cost-effectiveness of various interventions for newly diagnosed diabetic macular edema. Ophthalmology. 2013;120(9):1835–1842. doi:10.1016/j.ophtha.2013.02.002 [CrossRef]
  14. Martin DF, Maguire MG. Treatment choice for diabetic macular edema. N Engl J Med. 2015;372(13):1260–1261. doi:10.1056/NEJMe1500351 [CrossRef]
  15. Wells JA, Glassman AR, Jampol LM. Targeting the effect of VEGF in diabetic macular edema. N Engl J Med. 2015;373(5):481–482.

Respondent Characteristics

PatientSelf
Survey 1n = 199 (%)n = 211 (%)

“What is your practice setting?”Academic42 (21.11)52 (24.64)
Solo private practice24 (12.06)23 (10.90)
Group retina only practice82 (41.21)88 (41.71)
Multispecialty practice46 (23.12)46 (21.80)
Military2 (1.01)0 (0.00)
Other3 (1.51)2 (0.95)
“Where do you practice?”Northeast68 (34.17)69 (32.70)
South52 (26.13)60 (28.44)
Midwest44 (22.11)47 (22.27)
West coast35 (17.59)35 (16.59)
“How long have you been in practice?”1–10 years48 (24.12)54 (25.59)
11–20 years55 (27.64)78 (36.97)
21–30 years63 (31.66)57 (27.01)
31–40 years33 (16.58)22 (10.43)
“What is your gender?”Male183 (91.96)189 (89.57)
Female16 (8.04)22 (10.43)

Survey 2n = 107 (%)n = 109 (%)

“What is your practice setting?”Academic22 (20.56)52 (24.64)
Solo private practice6 (5.61)23 (10.90)
Group retina only practice48 (44.86)88 (41.71)
Multispecialty practice30 (28.04)46 (21.80)
Military0 (0.00)0 (0.00)
Other1 (0.93)2 (0.95)
“Where do you practice?”Northeast33 (30.84)69 (32.70)
South33 (30.84)60 (28.44)
Midwest21 (19.63)47 (22.27)
West coast20 (18.69)35 (16.59)
“How long have you been in practice?”1–10 years21 (19.63)54 (25.59)
11–20 years29 (27.10)78 (36.97)
21–30 years40 (37.38)57 (27.01)
31–40 years17 (15.89)22 (10.43)
“What is your gender?”Male101 (94.39)189 (89.57)
Female6 (5.61)22 (10.43)

Bevacizumab for “Self” Versus “Patient” in Survey 1

Patient DescriptionX2P ValueOdds Ratio (95% CI)
Phakic, 35 years old, and has BDR< 0.00010.268 to 0.610
Phakic, 35 years old, and has non-high-risk PDR< 0.00010.278 to 0.632
Phakic, 35 years old, and has moderate glaucoma and BDR0.00010.301 to 0.680
Phakic, 65 years old, and has BDR0.00010.296 to 0.674
Phakic, 65 years old, and has non-high-risk PDR0.00030.309 to 0.703
Phakic, 65 years old, and has moderate glaucoma and BDR0.00020.304 to 0.690
Pseudophakic, 65 years old, and has BDR< 0.00010.284 to 0.649
Pseudophakic, 65 years old, and has non-high-risk PDR0.00020.301 to 0.685
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.00010.296 to 0.672
Phakic and has not responded to anti-VEGF therapy0.68090.295 to 6.473
Pseudophakic and has not responded to anti-VEGF therapy0.20740.449 to 40.14

Aflibercept for “Self” Versus “Patient” in Survey 1

Patient DescriptionX2P ValueOdds Ratio (95% CI)
Phakic, 35 years old, and has BDR< 0.00012.079 to 5.678
Phakic, 35 years old, and has non-high-risk PDR< 0.00011.978 to 5.316
Phakic, 35 years old, and has moderate glaucoma and BDR< 0.00012.136 to 5.674
Phakic, 65 years old, and has BDR< 0.00011.931 to 5.210
Phakic, 65 years old, and has non-high-risk PDR< 0.00011.927 to 5.108
Phakic, 65 years old, and has moderate glaucoma and BDR< 0.00012.004 to 5.230
Pseudophakic, 65 years old, and has BDR< 0.00011.721 to 4.775
Pseudophakic, 65 years old, and has non-high-risk PDR< 0.00012.117 to 5.835
Pseudophakic, 65 years old, and has moderate glaucoma and BDR< 0.00012.022 to 5.257
Phakic and has not responded to anti-VEGF therapy0.56740.388 to 1.681
Pseudophakic and has not responded to anti-VEGF therapy0.62710.273 to 2.189

Geographic Distribution of Bevacizumab Use in Survey 1

Patient DescriptionX2P ValueOdds Ratio (95% CI), Northeast vs. SouthOdds Ratio (95% CI), Northeast vs. MidwestOdds Ratio (95% CI), Northeast vs. West Coast
Phakic, 35 years old, and has BDR0.00091.435 to 4.1581.302 to 3.9841.514 to 5.158
Phakic, 35 years old, and has non-high-risk PDR0.00041.655 to 4.8231.282 to 3.9221.471 to 4.987
Phakic, 35 years old, and has moderate glaucoma and BDR0.00121.519 to 4.3941.016 to 3.0861.462 to 4.930
Phakic, 65 years old, and has BDR0.00041.430 to 4.5131.266 to 3.8761.758 to 6.015
Phakic, 65 years old, and has non-high-risk PDR0.00021.592 to 4.6371.307 to 4.0001.718 to 5.855
Phakic, 65 years old, and has moderate glaucoma and BDR0.00071.436 to 4.1671.156 to 3.5341.671 to 5.686
Pseudophakic, 65 years old, and has BDR0.00071.339 to 3.9151.195 to 3.6901.780 to 6.109
Pseudophakic, 65 years old, and has non-high-risk PDR0.00031.422 to 4.1601.468 to 4.5451.731 to 5.920
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.00151.437 to 4.1741.212 to 3.7041.484 to 5.025
Phakic and has not responded to anti-VEGF therapy0.94010.285 to 11.3290.206 to 11.364<0.001 to >999.999
Pseudophakic and has not responded to anti-VEGF therapy0.88150.061 to 8.1570.238 to 13.699<0.001 to >999.999

Bevacizumab for “Self” Versus “Patient” in Survey 2

VA 20/40 or BetterVA 20/50 or Worse
Patient DescriptionX2P ValueOdds Ratio (95% CI)X2P ValueOdds Ratio (95% CI)
Phakic, 35 years old, and has BDR0.11620.892 to 2.8240.20840.790 to 2.950
Phakic, 35 years old, and has non-high-risk PDR0.07890.942 to 2.9940.13440.855 to 3.236
Phakic, 35 years old, and has moderate glaucoma and BDR0.04681.008 to 3.2360.23230.769 to 2.959
Phakic, 65 years old, and has BDR0.30580.759 to 2.4150.28900.740 to 2.747
Phakic, 65 years old, and has non-high-risk PDR0.06870.960 to 3.0670.19270.798 to 3.067
Phakic, 65 years old, and has moderate glaucoma and BDR0.26240.781 to 2.4810.28190.737 to 2.849
Pseudophakic, 65 years old, and has BDR0.05280.993 to 3.2360.16850.816 to 3.205
Pseudophakic, 65 years old, and has non-high-risk PDR0.06200.973 to 3.1450.07550.938 to 3.704
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.41130.715 to 2.2730.36520.694 to 2.695
Phakic and has not responded to anti-VEGF therapy0.39280.264 to 29.4120.15340.546 to 47.619
Pseudophakic and has not responded to anti-VEGF therapy0.31450.319 to 34.4830.96910.059 to 18.868

Geographic Distribution of Bevacizumab Use in Survey 2

VA 20/40 or better

Patient DescriptionX2P ValueOdds Ratio (95% CI), Northeast vs. SouthOdds Ratio (95% CI), Northeast vs. MidwestOdds Ratio (95% CI), Northeast vs. West Coast

Phakic, 35 years old, and has BDR0.01580.549 to 2.5091.524 to 7.2990.736 to 3.823
Phakic, 35 years old, and has non-high-risk PDR0.01620.579 to 2.7471.524 to 7.3530.879 to 4.586
Phakic, 35 years old, and has moderate glaucoma and BDR0.00620.684 to 3.1671.815 to 8.9290.852 to 4.516
Phakic, 65 years old, and has BDR0.01730.730 to 3.4501.567 to 7.6340.983 to 5.292
Phakic, 65 years old, and has non-high-risk PDR0.02420.692 to 3.2441.499 to 7.2990.905 to 4.820
Phakic, 65 years old, and has moderate glaucoma and BDR0.02060.632 to 2.9701.486 to 7.1430.922 to 4.901
Pseudophakic, 65 years old, and has BDR0.02890.727 to 3.5381.511 to 7.5190.828 to 4.624
Pseudophakic, 65 years old, and has non-high-risk PDR0.04770.618 to 2.9791.272 to 6.2110.971 to 5.249
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.02250.681 to 3.2391.473 to 7.1430.996 to 5.354
Phakic and has not responded to anti-VEGF therapy0.6302< 0.001 to > 999.999< 0.001 to > 999.999< 0.001 to > 999.999
Pseudophakic and has not responded to anti-VEGF therapy0.9978< 0.001 to > 999.999< 0.001 to > 999.999< 0.001 to > 999.999

VA 20/50 or worse

Patient DescriptionX2P ValueOdds Ratio (95% CI), Northeast vs. SouthOdds Ratio (95% CI), Northeast vs. MidwestOdds Ratio (95% CI), Northeast vs. West Coast

Phakic, 35 years old, and has BDR0.77920.467 to 2.7510.524 to 3.1850.630 to 4.180
Phakic, 35 years old, and has non-high-risk PDR0.74690.411 to 2.4840.518 to 3.1550.618 to 4.105
Phakic, 35 years old, and has moderate glaucoma and BDR0.80170.439 to 2.7260.550 to 3.4480.580 to 4.098
Phakic, 65 years old, and has BDR0.72990.422 to 2.5260.577 to 3.4010.607 to 3.997
Phakic, 65 years old, and has non-high-risk PDR0.45540.460 to 2.9770.712 to 4.4250.732 to 5.058
Phakic, 65 years old, and has moderate glaucoma and BDR0.71340.500 to 3.1710.613 to 3.9060.630 to 4.525
Pseudophakic, 65 years old, and has BDR0.45290.396 to 2.6410.640 to 4.0650.724 to 5.004
Pseudophakic, 65 years old, and has non-high-risk PDR0.43600.569 to 3.7320.717 to 4.7620.805 to 5.783
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.77180.505 to 3.1960.544 to 3.5460.628 to 4.491
Phakic and has not responded to anti-VEGF therapy0.5759< 0.001 to > 999.999< 0.001 to > 999.999< 0.001 to > 999.999
Pseudophakic and has not responded to anti-VEGF therapy0.9718< 0.001 to > 999.999< 0.001 to > 999.999< 0.001 to > 999.999

Aflibercept for “Self” Versus “Patient” in Survey 2

VA 20/40 or betterVA 20/50 or worse
Patient DescriptionX2P ValueOdds Ratio (95% CI)X2P ValueOdds Ratio (95% CI)
Phakic, 35 years old, and has BDR0.00021.792 to 6.2710.04221.022 to 3.411
Phakic, 35 years old, and has non-high-risk PDR0.00041.650 to 5.6780.02161.109 to 3.698
Phakic, 35 years old, and has moderate glaucoma and BDR< 0.00012.056 to 7.2440.02641.085 to 3.696
Phakic, 65 years old, and has BDR0.00011.804 to 6.2460.05660.984 to 3.185
Phakic, 65 years old, and has non-high-risk PDR0.00011.795 to 6.2120.04821.005 to 3.285
Phakic, 65 years old, and has moderate glaucoma and BDR< 0.00011.932 to 6.6830.01131.193 to 4.001
Pseudophakic, 65 years old, and has BDR0.00061.631 to 5.9240.07870.943 to 2.957
Pseudophakic, 65 years old, and has non-high-risk PDR< 0.00011.994 to 7.1510.03391.049 to 3.344
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.00031.664 to 5.6230.01331.175 to 3.993
Phakic and has not responded to anti-VEGF therapy0.39860.241 to 1.7630.77320.310 to 2.391
Pseudophakic and has not responded to anti-VEGF therapy0.34590.069 to 2.5530.28740.109 to 1.931

Ranibizumab for “Self” Versus “Patient” in Survey 2

VA 20/40 or betterVA 20/50 or worse
Patient DescriptionX2P ValueOdds Ratio (95% CI)X2P ValueOdds Ratio (95% CI)
Phakic, 35 years old, and has BDR0.00020.185 to 0.7760.04200.066 to 0.951
Phakic, 35 years old, and has non-high-risk PDR0.00040.224 to 0.9160.10440.133 to 1.208
Phakic, 35 years old, and has moderate glaucoma and BDR< 0.00010.173 to 0.7480.03550.064 to 0.908
Phakic, 65 years old, and has BDR0.00010.145 to 0.6100.06550.099 to 1.074
Phakic, 65 years old, and has non-high-risk PDR0.00010.201 to 0.8120.09700.131 to 1.183
Phakic, 65 years old, and has moderate glaucoma and BDR< 0.00010.140 to 0.5880.01140.071 to 0.714
Pseudophakic, 65 years old, and has BDR0.00060.175 to 0.7200.02410.059 to 0.821
Pseudophakic, 65 years old, and has non-high-risk PDR< 0.00010.156 to 0.6520.03670.108 to 0.932
Pseudophakic, 65 years old, and has moderate glaucoma and BDR0.00030.155 to 0.6330.01140.051 to 0.684
Phakic and has not responded to anti-VEGF therapy0.3986< 0.001 to > 999.9990.82700.063 to 9.163
Pseudophakic and has not responded to anti-VEGF therapy0.3459< 0.001 to > 999.9990.79470.084 to 25.385
Authors

From Rutgers-Robert Wood Johnson Medical School, Piscataway, NJ (CGM, GB, KWJ, HFF, DR, JLP); and NJ Retina, New Brunswick, NJ (HFF, DR, JLP).

Disclosures: Dr. Fine is a consultant and/or speaker for Allergan, Genentech, and Regeneron. Dr. Prenner is a consultant for Regeneron, Ophthotech, Panoptica, and Alcon, as well as an equity owner for Ophthotech and Panoptica. The remaining authors report no relevant financial disclosures.

Address correspondence to Jonathan L. Prenner, MD, 10 Plum Street, 6th Floor, NJ Retina, Rutgers, Robert Wood Johnson Medical School, New Brunswick, NJ 08901; email: jonathanprenner@gmail.com.

Received: September 27, 2015
Accepted: March 24, 2016

10.3928/23258160-20160601-06

Sign up to receive

Journal E-contents