RT Journal Article SR Electronic T1 Retina specialists treating cystoid macular oedema secondary to retinal vein occlusion recommend different treatments for patients than they would choose for themselves JF British Journal of Ophthalmology JO Br J Ophthalmol FD BMJ Publishing Group Ltd. SP 1332 OP 1336 DO 10.1136/bjophthalmol-2015-307849 VO 100 IS 10 A1 Wang, Marlene D A1 Jeng-Miller, Karen W A1 Feng, Henry L A1 Prenner, Jonathan L A1 Fine, Howard F A1 Shah, Sumit P YR 2016 UL http://bjo.bmj.com/content/100/10/1332.abstract AB Aims To evaluate the presence of cognitive bias among retinal specialists when recommending treatment options for cystoid macular oedema (CMO) secondary to retinal vein occlusion (RVO).Methods Two randomly chosen samples of retina specialists were surveyed regarding their treatment and dosing regimen choices among three antivascular endothelial growth factor (anti-VEGF) biologics (aflibercept, bevacizumab and ranibizumab), intravitreal steroid, focal laser and observation for the treatment of CMO secondary to RVO. The first group was asked to make recommendations for two hypothetical patients: one with CMO secondary to branch RVO (BRVO) and the second with CMO secondary to central RVO (CRVO). The second group was asked to make recommendations as if they themselves were the hypothetical patient with the same disease processes.Results The survey was completed by 492 respondents (20.1%). When comparing anti-VEGF agents for patients with BRVO, a majority of physicians recommended bevacizumab (60.5%) over ranibizumab (37.8%) and aflibercept (1.7%; p<0.0001). For themselves, physicians were more likely to recommend ranibizumab (44.9%) over bevacizumab (39.2%) and aflibercept (15.9%; p<0.0001). When comparing among the anti-VEGF agents chosen for patients with CRVO, a majority of physicians recommended bevacizumab (56.7%) over ranibizumab (28.2%) and aflibercept (15.1%; p<0.0001), but when choosing for themselves, retina specialists were equally divided among the three biologics (aflibercept 30.6%, bevacizumab 36.5% and ranibizumab 32.9%; p=0.559). The results were influenced by geographical location but not by the gender, the length of practice or the type of practice.Conclusions Physicians should be aware that cognitive biases exist and take this into consideration when making treatment recommendations for their patients.