Physician and Patient Knowledge of Safety and Safe Use Information for Aflibercept in Europe: Evaluation of Risk-Minimization Measures

The anti-vascular endothelial growth factor (VEGF)-A therapy aflibercept (Eylea; Bayer AG), administered as an intravitreal injection, is approved by the European Medicines Agency in adults for the treatment of neovascular (wet) age-related macular degeneration (wAMD), visual impairment due to macular edema secondary to central retinal vein occlusion (CRVO), visual impairment due to macular edema secondary to branch retinal vein occlusion (BRVO), visual impairment due to diabetic macular edema (DME), and visual impairment due to myopic choroidal neovascularization [1].

Intravitreal injections, including anti-VEGF therapies, have been associated with ocular complications including endophthalmitis, transient increases in intraocular pressure, traumatic cataract, and retinal and vitreous detachment. Less serious and more common complications include conjunctival hemorrhage, vitreous floaters, and eye pain [2, 3].

A risk-management plan is in place for aflibercept in the European Union [4]. As part of the risk-management plan, the manufacturer has developed materials to educate physicians and patients on the key safety information and safe use for aflibercept. The key content of the educational materials for physicians includes the importance of using the correct sterile injection technique and monitoring and managing potential injection-related adverse events. Key content for patients includes steps for preparing for treatment, monitoring for adverse events, and steps to take if they identify adverse events.

The objectives of this study were to investigate whether physicians and patients received the educational materials and to assess physicians’ and patients’ knowledge and understanding of key safety information therein. A similar study was conducted to evaluate physician knowledge of the key safety information for aflibercept in Canada [20].

Few post-authorization safety studies have described physicians’ reported receipt of educational materials in the published literature [9‐12]. In only one case was reported receipt as high as 75% [10]. Other studies have reported ranges from 37 to 51% [9, 11, 12]. Moreover, a survey of 800 European physicians found that reported receipt of educational materials ranged from 16% to 69% across participating countries [12]. In our study, most physicians reported receipt of the summary of product characteristics (87%) and prescriber guide (77%). Approximately half reported receipt of the intravitreal injection procedure video (50%) and the patient booklet (54%).

Patient-reported receipt of the educational materials, in contrast, was relatively low (23–38% for individual materials), potentially reflecting poor recall if the materials had indeed been received. Although direct comparison cannot be made between physician and patient assessments because they involved two independent samples, in general, there was a disconnect between the proportion of physicians reporting receipt of the patient booklet (54%) and the proportion of patients reporting receipt of the patient booklet (38%). A potential driver for this disconnect is that physicians in some countries (e.g., France and Spain) are required to have patients sign an informed consent form with relevant safety information prior to treatment; therefore, it is possible that some physicians could prioritize competing information sources that are legally required over the patient booklet. Despite variability in respondents’ receipt of the educational material, it was encouraging that, among physicians and patients who reported having received the material, use was high.

Published information describing acceptable levels of knowledge and behavior related to risk-minimization measures is somewhat limited. Knox and colleagues [13] reported on patient understanding of medication guides from a review of 66 assessment reports submitted to regulatory authorities. Only 30% of studies achieved an 80% knowledge level (% correct) for the most important risk communicated in the medication guide, and the mean knowledge level was 63.8%. A review of risk-minimization studies posted on the European Union electronic Register of Post-Authorisation Studies showed high variability in measures of program success criteria, variability in the nature of reporting, and variability of results, with approximately half of studies reporting having been “successful” but a large percentage in which results were inconclusive (17%) or were not adequately reported to support a conclusion (31%) [14].

With respect to our study, some of the most important information communicated in the aflibercept educational materials is related to side effects. We found that over 80% of physicians responded correctly to most questions on this topic. Physician knowledge varied across the other information categories. In general, physicians’ knowledge of storage and preparation guidelines, safe use, and injection procedures was also high. Knowledge on dosing guidelines (with the exception of monitoring requirements) varied by indication (61–94%), perhaps because some indications were approved recently or are encountered infrequently in clinical practice or because participating physicians prescribed aflibercept for these indications less frequently than for other indications. In general, physicians demonstrated a conservative approach to the monitoring questions, with a minority correctly indicating that there is no requirement for monitoring between injections for wAMD (28%) or DME (26%). Physician knowledge was lower on questions related to aflibercept use in pregnancy (with 48–59% answering these questions correctly); however, in several cases, physicians who answered incorrectly selected another response reflecting a more conservative treatment approach. These findings are largely consistent with those from a study of Canadian physicians’ knowledge of safe use for aflibercept, in which 60% of physicians knew that aflibercept should not be used in pregnancy unless clearly indicated by medical need in which benefits outweigh risks and 21% responded more conservatively that aflibercept should never be used in pregnancy [20].

Although patient knowledge was high for health conditions they should tell their ophthalmologist about before having an aflibercept injection (85–92%), approximately half of patients overall, and proportionally more women than men, correctly responded to the question related to pregnancy and breastfeeding. This is likely an effect of the age of the aflibercept patient population; 99% of the patients in the study were aged 46 years or older. Patient knowledge was high for side effects that are easier to identify but was lower for more complex side effects (e.g., “detachment of the gel-like substance inside the eye from the retina,” 42%). Most patients (78%) knew that they should contact their ophthalmologist’s clinic immediately if they thought they might be experiencing a side effect.

A key strength of the study is the diversity of physician and patient participants. The targeted numbers of physician and patient respondents were achieved. Physicians were well distributed by their ophthalmology specialty, sex, years treating patients, and aflibercept prescribing practices. The characteristics of patients by demographics and their aflibercept treatment history were also diverse. An additional strength is that the physician assessment was conducted after physicians had received the educational material and had an opportunity to use them in clinical practice, and the patient assessment was conducted after patients had received aflibercept and had the opportunity to receive the educational material. Moreover, the completion rate for the patient assessment was high [773 patients completed the questionnaire among 851 who were invited and eligible (91%)] and exceeded the completion rate among physicians (5.1%). The favorable response rate among patients may have occurred in part because patients were recruited in person by a trusted healthcare professional. Recruitment and participation happened at their doctor’s office as part of a regularly scheduled visit; an additional clinic visit was not required. Further, the questionnaire was brief, with 12 questions, and took approximately 10–15 min to complete, thus minimizing the burden on the patient and the site. An additional study strength is that the accuracy of responses among physicians and patients was facilitated by formal cognitive pretesting of the questionnaires.

The results of this study must be considered with several limitations in mind. Although the study was designed to ensure the selection of a diverse and generally representative sample of prescribers and patients, there was no exhaustive list of all aflibercept prescribers and patients from which to draw a random sample. Physician response rates to surveys have been reported to be low in general [15, 16], and particularly low in studies measuring the effectiveness of risk-minimization measures in Europe [9‐11]. A low response rate among physicians could result in bias. Impediments to participation, including time, regulatory requirements, and reporting and contract requirements, have been well described by Madison et al. [17]. In addition, the study did not involve any countries in Eastern Europe, potentially limiting its external validity. Disproportionately more patients in our sample did not have a university-level education compared with the general population aged 55–74 years in Europe [18]. In general, study participants may not necessarily represent all relevant prescribers or patients. Moreover, respondents who completed the questionnaire may have differed from non-respondents in characteristics measured in the questionnaire (e.g., knowledge and use of the materials). The direction and magnitude of such potential respondent bias are not known. Another potential limitation of the patient assessment is that the study could influence sites to provide more education to patients than they ordinarily would. Nevertheless, to mitigate potential biases and an “intervention effect,” site personnel were trained to provide only limited information about the study before patients’ participation and to practice objective interviewing techniques, and patients completed the questionnaire at the site before receiving any additional counseling about treatment. Finally, according to Banerjee and colleagues’ hierarchical framework to assess the effectiveness of risk-minimization measures [19], this study does not explore the most valuable endpoints of prescriber behavior and adverse health outcomes, which is an additional limitation, and therefore cannot measure the ultimate success of the risk-minimization measures.

Most physicians reported receiving the summary of product characteristics (87%), prescriber guide (77%), and patient booklet (54%); 50% reported receiving the intravitreal injection procedure video. Patient receipt of the educational material was variable (patient booklet: 38%; audio CD: 23%; patient leaflet: 35%). The observed patterns of knowledge among physicians and patients indicated the greatest knowledge of the most important risks emphasized in the educational material and other product information and lower knowledge of more complex aspects of safe use (e.g., concepts related to dosing and monitoring), for which we would assume that physicians would consult the label and/or prescriber guide rather than relying on recall. Likewise, levels of patient knowledge indicated the greatest knowledge of less complex concepts and lower knowledge of more complex concepts and issues less salient to the patient population. Relatively high levels of patient knowledge despite low reported receipt of the educational materials suggests that patients were receiving information from other sources.