Article Text
PDF
Abstract
Background To compare individual ophthalmologists grouped as outliers or non-outliers based on the mean 12-month visual acuity (VA) outcomes for their patients with neovascular age-related macular degeneration (nAMD).
Methods This prospectively designed database study included treatment-naïve eyes with nAMD starting vascular endothelial growth factor inhibitors between July 2018 and April 2023 in Australia. Ophthalmologists were classified into high outliers, non-outliers and low outliers with a funnel plot of the adjusted mean 12-month VA change. The number of injections, last injection interval and proportion of visits where choroidal neovascularisation was active were compared between the groups.
Results 38 ophthalmologists who treated a total of 1266 eyes (male, 35%; mean age, 81 years old) were classified into 1 high outlier, 34 non-outliers and 3 low outliers (mean VA change, 7.5, 5.1 and 2.5 letters, respectively). The high outlier gave significantly more injections than the non-outliers (mean, 8.6 vs 7.7; p<0.001), while the low outliers administered significantly fewer injections than the non-outliers (mean, 7.1 vs 7.7; p=0.009). The last injection interval was shortest in the high outlier’s eyes (9.4 weeks), followed by non-outliers’ (10.8 weeks; p=0.04 (vs high outlier’s)) and low outliers’ (11.8 weeks; p=0.22 (vs non-outliers’)). The low outliers’ patients had more visits with intraretinal fluid (59%) than non-outliers’ (29%; p<0.001) and high outlier’s patients (31%; p<0.001).
Conclusion The low outliers’ eyes had fewer injections, a longer treatment interval and more visits with intraretinal fluid. Building a system through which low outliers are anonymously notified of their performance would help improve general quality of care.
- Age-Related Macular Degeneration
- Epidemiology
Data availability statement
No data are available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
More injections generally deliver better outcomes, but there are limited data on whether undertreatment occurs at the level of ophthalmologists.
WHAT THIS STUDY ADDS
Ophthalmologists who have outlying results can be identified through the funnel plot and are associated with worse outcomes.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Detecting outlier ophthalmologists is the first step in a quality assurance process whereby low outliers are notified of their status.
Introduction
Clinical trials and real-world data have unequivocally established that vascular endothelial growth factor (VEGF) inhibitors are the standard of care for the treatment of neovascular age-related macular degeneration (nAMD).1–3 The next step in improving patient visual outcomes is to compare the performance (visual outcomes) of different ophthalmologists in clinical quality registries.4
This comparison can lead to ophthalmologists learning from each other; specifically, underperforming ophthalmologists can be alerted and motivated to improve their practice when they can see a comparison of their own patient outcomes with those of their peers.4 Barriers to delivery of the care at the patient level, such as out-of-pocket costs and remote locations with limited transport,5 6 may be identified. This self-improving health system reduces the risk of harm and improves the outcomes of healthcare.4
The Fight Retinal Blindness! (FRB!) registry was prospectively designed to collect large amounts of high-quality data on treatment outcomes of eyes with nAMD from multiple countries.7 This allows us to identify underperforming and overperforming ophthalmologists in terms of their patients’ visual outcomes and to investigate the difference in treatment delivered such as the number of injections per year. Considering that more injections are likely to be associated with better visual outcomes,8–10 the finding and investigation of significant differences in treatment delivered between ophthalmologists, if any, may help improve the visual outcomes in the underperforming ophthalmologists.
Therefore, the current study aimed to detect outlier ophthalmologists based on the performance of 12-month visual outcomes for nAMD and to investigate any difference in treatments given and visit characteristics between outliers and non-outliers.
Materials and methods
Data source
This is a prospectively designed multicentre cohort study using the FRB! registry of real-world treatment outcomes worldwide,11 which enables ophthalmologists to perform self-audit and research. Ophthalmologists commit to tracking at least 85% of eyes with the relevant condition in a variety of available modules in the FRB! project, including the nAMD module. Patients and treating physicians made all management decisions, thereby reflecting real-world practice. Urban and regional practices are included in this registry. Over 90% of intravitreal therapy is provided through private practices in Australia (figure 1).
Distribution of practices in Australia participating in the Fight Retinal Blindness! registry.
The FRB! registry collects a prespecified, minimum dataset which ensures that data from each visit are 100% complete and within prespecified ranges. Fields collected at baseline visits include gender, birth year, pretreatments (or treatment naïve), the first treatment and baseline visual acuity (VA; best of uncorrected, corrected or pinhole). Minimum data collected at each follow-up visit include VA, treatments given (aflibercept, ranibizumab or bevacizumab, etc), choroidal neovascularisation (CNV) activity, presence of macular atrophy (MA) and subretinal fibrosis (SRFi) and adverse events. The CNV activity was classified into ‘inactive’, ‘active, (subretinal fluid [SRF] only)’ hereafter called ‘SRF-only’ or ‘active (any combination of SRF, intraretinal fluid [IRF] or haemorrhage, ie, excluding SRF-only)’ hereafter called ‘IRF’.11 12
Eye selection
We identified treatment-naïve eyes with nAMD in the FRB! registry that started VEGF inhibitors (aflibercept, ranibizumab or bevacizumab) in Australia between July 2018 and April 2023, when detailed information on CNV activity (SRF-only or IRF), MA or SRFi was available. Patients treated in Australia were only included to exclude the effect of different healthcare systems and treatment regimens among multiple countries. We excluded eyes that (a) did not have age, gender or VA data at baseline, (b) CNV was inactive at baseline, (c) did not have a 12-month visit (365±90 days) and (d) were treated by ophthalmologists with ≤4 eligible eyes (figure 2). Follow-up of eyes started when the first VEGF inhibitor was administered and ended at the closest date to 12 months.
Selection of eyes. CNV, choroidal neovascularisation; VEGF, vascular endothelial growth factor.
Performance indicators and case-mix adjustment
The primary performance indicator was the differences between observed (crude) and expected (adjusted) values in mean 12-month VA change of all the eyes that an individual ophthalmologist had treated.13 14 We used linear models adjusted by age, gender, VA, CNV activity at baseline (SRF-only or IRF), the initial treatment (aflibercept, bevacizumab or ranibizumab) and the existence of MA and SRFi at baseline to calculate the expected 12-month VA change for each eye considering the case-mix differences between the ophthalmologists.13 We calculated the difference between the expected and observed values, aggregating them by ophthalmologists to obtain the mean of the differences.
Funnel plots
Funnel plots allow graphical benchmark of the performance of doctors and hospitals, taking in to account random variation.4 15 The Australian framework for clinical quality registries recommends the level of control limits (CIs) should be 95% to identify outliers.4 We classified the ophthalmologists into three groups according to their positions relative to the 95% control limits in the funnel plot:
High outliers: falling outside the upper control limit.
Non-outliers: falling inside the control limits.
Low outliers: falling outside the lower control limit.
The mean of the differences between the observed and expected values was used as the indicator for the funnel plot. The 95% control limits were constructed with the non-parametric method advocated by Kuhrij et al because VA change is a continuous variable.16
Outcomes
The following parameters within 12 months were compared between the three groups:
Number of injections.
Last injection interval.
Number of visits.
Proportion of eyes that achieved lesion inactivation at least once.
Time to initial lesion inactivation.
Proportion of eyes that developed MA that did not have MA at baseline.
Proportion of eyes that developed SRFi that did not have SRFi at baseline.
Proportion of visits at which an injection was given.
Proportion of visits at which CNV was active (including the breakdown of SRF-only and IRF).
The difference and 95% CIs for (1)–(4), (6) and (7) were estimated using linear models in which each parameter was regressed on the three groups. Those for the (8) and (9) were estimated similarly except for the number of visits being used as weights. Multiplicity adjustments with the Tukey method were performed.17 The median time and 95% CI for (5) was calculated with a Kaplan-Meier estimator where eyes that continued to be active throughout the period were regarded as censored at their last visits around 1 year. Kaplan-Meier curves between the three groups were compared with log-rank tests considering multiplicity adjustments by the Benjamini-Hochberg method.18 We did not adjust for covariates at this stage because adjustment was applied when constructing the funnel plots, an approach described by Resnik and Hart.14
Before the COVID-19 pandemic
As the period of the COVID-19 lockdowns could have unpredictably influenced results, we similarly analysed the data in and before December 2019, as the first confirmed case of COVID-19 was found in January 2020 in Australia.19 Since the grading of CNV activity, MA or SRFi was not available during this period, we adjusted the VA change just with age, gender, VA and the initial treatment.
A p value <0.05 was considered statistically significant. All statistical analyses were conducted using R software V.4.2.3 (R Foundation for Statistical Computing).
Results
The selection of eyes is shown in figure 2. A total of 1266 eyes (male, 35%; mean age, 81 years old) from 38 ophthalmologists were eligible for this analysis (table 1). The mean VA at baseline for eyes overall was 62 letters (table 1).
Baseline characteristics of the high outliers, non-outliers and low outliers classified by the funnel plot by 95% control limits
Funnel plot
Figure 3 shows a funnel plot with 95% control limits on the difference between the observed and expected values of mean VA change at 12 months for each practitioner. One ophthalmologist fell above the upper control limit (high outlier), three below the lower control limit (low outliers) and the remaining 34 ophthalmologists fell between the control limits (non-outliers). The baseline characteristics of the eyes treated by the high outlier, non-outliers and low outliers are shown in table 1, with the number of eyes per ophthalmologist being 173, 22 and 31 eyes and mean changes in VA at 12 months being 7.5, 5.1 and 2.5 letters, respectively (table 2).
Funnel plot of change in visual acuity at 12 months by practitioners using 95% control limits. Each dot represents one ophthalmologist (orange, high outliers; grey, non-outliers; blue, low outliers). The x-axis represents the expected number of eyes adjusted by the case-mix factors at baseline such as age, gender, visual acuity, the initial type of vascular endothelial growth factor inhibitors, fluid type at baseline (subretinal fluid-only or intraretinal fluid) and the existence of subretinal fibrosis and macular atrophy. The y-axis represents the difference between the observed and expected values. The central dot lines represent the reference standard. The 95% limits are constructed by the non-parametric method.16
Treatments delivered and the proportion of visits when CNV was active during 12 months for the high outliers, non-outliers and low outliers
Supplemental material
Treatments delivered
The high outlier administered significantly more injections over the 12 months than the non-outliers (mean, 8.6 vs 7.7; p<0.001), while the low outliers administered significantly fewer injections than the non-outliers (mean, 7.1 vs 7.7; p=0.009) (table 2). In keeping with this difference, the last interval of the injections for the high outlier’s eyes was significantly shorter than the non-outliers’ (mean, 9.4 weeks vs 10.8 weeks; p=0.04), whereas the interval was not significantly different between the low outliers and non-outliers’ eyes (mean, 11.8 weeks vs 10.8 weeks; p=0.22). The high outlier’s eyes had significantly more visits than the non-outliers’ (mean, 8.7 vs 8.0; p<0.001), while the low outliers’ eyes had significantly fewer visits than the non-outliers’ (mean, 7.3 vs 8.0; p<0.001). The proportion of visits at which an injection was given was lowest in the non-outliers, but all of them were above 95% (99%, 99% and 96% for the high outlier, non-outliers and low outliers, respectively).
CNV activity by outlier status
Significantly fewer eyes treated by the low outliers achieved the lesion inactivation than those by the non-outliers (86% vs 92%; p=0.02) and the high outlier (86% vs 95%; p=0.005). The median time to initial inactivation for overall eyes was 8.1 weeks. The low outliers’ eyes had longer median time to initial inactivation than the non-outliers’ (20.0 weeks vs 8.0 weeks) (table 2) with significantly different Kaplan-Meier curves (p<0.001) (online supplemental figure 1), whereas that time for the high outlier and non-outliers’ were similar (8.7 weeks vs 8.0 weeks) with similar Kaplan-Meier curves (p=0.62) (online supplemental figure 1).
The low outliers’ eyes had a significantly higher proportion of visits at which CNV was active than the non-outliers’ (70% vs 48%; p<0.001), whereas this was similar between the high outlier and non-outliers (49% vs 48%; p=0.92) (table 2). The low outliers’ eyes had a significantly lower proportion of SRF-only visits than the non-outliers’ (11% vs 19%; p=0.002) and a significantly higher proportion of IRF visits than the non-outliers’ (59% vs 29%; p<0.001). On the other hand, the high outlier’s patients had similar proportions of SRF-only visits and IRF visits to non-outliers (SRF-only, 24% vs 27%; p=0.72; IRF, 23% vs 23%; p=0.40).
Development of MA and SRFi
The same proportion of eyes without MA at baseline developed MA within 12 months: 11% (17/153), 11% (90/804) and 11% (14/128) for the high outlier, non-outliers and low outliers’ eyes, respectively. Of the eyes without SRFi at baseline, 7.5% (11/147) developed SRFi for the high outlier, 8.8% (70/791) for the non-outliers and 13% (15/113) for the low outliers, which did not show significant differences between any groups.
Before the COVID-19 pandemic
A total of 45 ophthalmologists were included (online supplemental table 1), of which 5, 37 and 3 ophthalmologists were classified into high outliers, non-outliers and low outliers, respectively, with a mean VA change of 9.7, 5.0 and 1.1 letters (online supplemental figure 2 and table 2). Baseline characteristics are shown in online supplemental table 1. The high outliers and non-outliers gave a similar mean number of injections (8.2 vs 8.3; p=0.88), with a similar last injection interval (9.2 weeks vs 9.7 weeks; p=0.63) and number of visits (8.6 vs 8.7; p=0.86) (online supplemental table 2). Meanwhile, the low outliers’ eyes received significantly fewer injections than non-outliers’ eyes (7.8 vs 8.3; p=0.003) and had fewer visits (8.0 vs 8.6; p<0.001). The last injection interval was similar between these two groups (9.7 weeks vs 10.3 weeks; p=0.18). The low outliers’ eyes had longer median time to initial inactivation than the non-outliers’ with significantly different Kaplan-Meier curves (p<0.001) (online supplemental figure 3).
Discussion
While more injections generally deliver better outcomes,8–10 there are limited data on whether undertreatment (or overtreatment) occurs systematically in individual practitioners. We have used funnel plots to analyse 12-month visual outcomes at the level of practitioners. While only 5% of practitioners will by definition be classified as outliers,20 we were able to identify a high outlier and several low outliers who gave their patients fewer injections at fewer visits. Distinguishing outlier ophthalmologists using 12-month VA change is appropriate since the number of injections in the first year has been reported to be more strongly associated with long-term visual outcomes than those in the second year and third year.21 While this analysis has shown that doctor-level undertreatment can be identified and is associated with worse outcomes, it is also the first step in a quality assurance process whereby low outliers are notified of their status.
Some randomised controlled trials (RCTs) have reported that treat and extend (TAE) regimens are non-inferior to fixed treatment regimens in the initial 12 months (the TReat and extEND [TREND] study, a 6.2-letter gain after 8.7 injections (baseline, 59.5 letters)22; the Treat-and-Extend Protocol in Patients with Wet Age-Related Macular Degeneration [TREX-AMD] study, a 10.5-letter gain after 10.1 injections (baseline, 61 letters)23). However, observation studies have reported that fewer injections are given in real-world settings. A study in Australia reported that 7.5 injections were given in 12 months with a 5-letter gain (baseline, 56.5 letters).10 Studies using the Intelligent Research in Sight Clinical Data (IRIS) registry have reported generally lower mean VA gains: one in 2018 described a 2.5-letter gain after 6.1 injections (baseline, 56.5 letters)24 and another in 2024 a 3-letter gain after 7.2 injections (baseline, 60 letters).25 Both the high outlier and the non-outliers in the present study gave more injections (8.6 and 7.7 injections, respectively) than these studies. Even the low outliers gave 7.1 injections, which is comparable to those in the IRIS registry,25 with similar baseline VA of 60.4 letters and a similar smaller VA gain of 2.5 letters. These results suggest that Australian ophthalmologists contributing to FRB! registry, even the low outliers, treated patients more intensively than those in the USA with better VA outcomes. The high proportion (>95%) of visits at which treatments were given in all groups indicates that the majority of FRB! ophthalmologists in Australia use TAE regimens. The high outlier treated more eyes than the non-outliers and low outliers per ophthalmologist. This suggests that the high outlier might be a retinal specialist and the other groups included some comprehensive ophthalmologists.
The high outlier gave most injections in this study, followed by non-outliers and low outliers, which is consistent with previous reports that services that give more injections generally deliver better VA outcomes.6 8–10 As expected, the low outliers’ eyes had the longest final treatment interval. Our data suggest that outcomes of eyes treated by low outliers might have been better if they had extended treatment intervals less aggressively, since their patients’ eyes had the far higher proportion (70%) of visits at which CNV was active compared with the high outlier and non-outliers’ (approximately 50%). This is the likely reason that we found a clinically meaningful mean difference of 5.0 letters26 between eyes treated by high outlier and low outliers.
Interestingly, the low outliers’ eyes had the lowest proportion of SRF-only visits and the highest proportion of IRF visits. Previous studies have reported that SRF-only was associated with better outcomes.26 27 Sánchez-Monroy et al reported that eyes that were graded predominantly as having SRF-only had a significantly lower risk of developing MA with similar 5-year vision outcomes as eyes that were predominantly inactive.12 The Tolerating Subretinal Fluid in Neovascular Age-Related Macular Degeneration Treated with Ranibizumab Using a Treat-and-Extend Regimen (FLUID) study reported that a TAE regimen tolerating SRF achieved VA comparable to that achieved by a regimen not tolerating SRF.26 In contrast, IRF has been reported to be associated with worse VA outcomes.28 29 Eyes treated by low outliers would likely have had better VA outcomes if they had been treated more aggressively to eliminate IRF.
Eyes treated by the low outliers had a lower chance of achieving lesion inactivation within 1 year and longer time to initial lesion inactivation (20 weeks) than those by the non-outliers and high outlier (around 8–9 weeks). Considering the low outliers gave fewer injections than the other groups, these results suggest that eyes treated less frequently needed more time to initial lesion inactivation, which is consistent with a previous study.30 The median time to initial inactivation for overall eyes was 8.1 weeks in our study. This is somewhat shorter than 15 weeks reported in the study in 201530 perhaps because the association between intensive treatments and better VA outcomes have been increasingly recognised among ophthalmologists recently and because that study involved only patients receiving ranibizumab or bevacizumab.
The risk of developing MA was similar among the three groups (11% for all groups). Observational data suggest the main risk factor for the development of MA is low lesion activity, with one study finding that the quartile of eyes with the most gradings of lesion inactivity was four times more likely to develop MA than the quartile with the fewest gradings of lesion inactivity.31 Some RCTs have reported that monthly regimens had a significantly higher risk of MA than pro re nata but number of injections is likely directly correlated with lesion inactivity which the RCTs did not specifically analyse.32 33 The risks of developing MA in our study were lower than the values reported by the The Development of New Macular Atrophy In Patients With Neovascular Age-Related Macular Degeneration: a Comparison of Ranibizumab and Aflibercept (RIVAL) study (18–20% developed MA after 12 months with around 10 injections by a relatively intensive TAE regimen).34
Although the difference was not statistically significant, the low outliers’ eyes had a higher risk of developing SRFi (13%) than the non-outliers (8.8%) and high outlier’s (7.5%), which correlated with the proportion of IRF visits. Several observational studies reported that high-activity lesions were associated with the development of SRFi,35 in particular IRF was associated SRFi.12 These results are consistent with the findings of the present study.
Undertreatment may be caused by various factors. Some patients may have difficulty receiving injections frequently enough due to financial constraints or other barriers to seeing ophthalmologists frequently such as physical disabilities, comorbidities (eg, mental problems) or living in remote locations with limited access to the point of care.5 6 Ophthalmologists might limit the frequency of injections for several reasons, for example, if they were worried about exacerbating MA or other adverse events, if they had too many patients to treat or if they had low expectations of VA gain.5 6
Funnel plots are recommended to monitor the quality of care.4 Comparison of the circumstances of high outliers and non-outliers (eg, low outliers have more patients with limited transport options) will clarify specific steps that can be taken to break down the barrier (eg, subsidised taxi fares). Low outliers are unlikely to realise that their patients’ outcomes are significantly inferior to those of their peers, so they need to be notified of this. A formal process has been described in some jurisdictions, especially in public hospitals in Australia where monitoring of outcomes is often mandatory.4 Since the FRB! project relies on the goodwill of contributing ophthalmologists, ophthalmologists must be notified of their performance in a completely anonymised way, otherwise they may be reluctant to contribute data.
Our study has several limitations. First, the data on eyes such as CNV type and CNV lesion size36 or those on patient demographics such as distance to clinics and the socioeconomic status were not available. Establishing the differences in these variables between outliers and non-outliers using external data such as health records would improve the specificity of our recommendations. Second, the international generalisability of the results might be limited because ophthalmologists in Australia were only included. The internal validity, however, is high because all ophthalmologists practised under the same healthcare scheme, which would reduce confounding caused by different healthcare systems in different countries. Third, the results might have been influenced by the COVID-19 pandemic, where examinations or injections may have been omitted or delayed.37 The sensitivity analysis using data before the pandemic found that the low outliers gave fewer injections than the non-outliers and the high outliers.
In conclusion, we found that eyes with nAMD treated by low outlier ophthalmologists had fewer injections, fewer visits and longer interval of injections than eyes treated by the high outlier and non-outliers using a large, prospectively designed registry with mandatory fields in Australia. Establishing a process through which ophthalmologists who have outlying results can be anonymously notified of their performance will help improve patient visual outcomes and general quality of care.
Data availability statement
No data are available.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and institutional ethics approval was obtained by the Sydney Local Health District Human Research Ethics Committee (ETH00956) and the Royal Australian and New Zealand College of Ophthalmologists Ethics Committee (16.09). The research described adhered to the National Statement on Ethical Conduct in Human Research guidelines. Opt-out informed consent was approved in Australia because this is an observational study of treatment outcomes in routine clinical practice with no active intervention.
Acknowledgments
Blink, Australian Capital Territory (Dr R Barry); Bundaberg Eye Clinic, Queensland (Dr I McLean); Cairns Eye Surgery, Queensland (Dr A Field); Canberra Hospital, Australian Capital Territory (Dr J Wells); Casey Eye Centre, Victoria (Dr K Michalova); Central Coast Eye Specialist, New South Wales (Dr S Young); Coastwide Eye Surgery, New South Wales (Dr R Ferrier); Croydon Eye Clinic, Victoria (Dr R Nazari); Doncaster Eye Center, Victoria (Dr L Chow); Dorset Consultant Center, Victoria (Dr H Steiner); Dr Alex Amini's Practice, Victoria (Dr A Amini); Dr Niladri Saha GHR, South Australia (Dr N Saha); Dr Niladri Saha MV, South Australia (Dr N Saha); Dr Phillip Windle, Queensland (Dr P Windle); Eye Associates, New South Wales (Dr M Gillies); Eye Doctors Mona Vale, New South Wales (Dr P Beaumont); Eye Specialists Greensborough, Victoria (Dr L Chow); Eye Surgeons Miranda, New South Wales (Dr A Hunt); Eye Wide Bay, Queensland (Dr Z Louw); EyeClinic Albury Wodonga, New South Wales (Dr A Luckie); Eyemedics (Wayville), South Australia (S Lake, Dr D Qatarneh); Focus Eye Centre, New South Wales (Dr P Berdoukas); Forest Eye Surgery, New South Wales (Dr D Robaei); Gladesville Eye Specialists, New South Wales (Dr S Young); Gordon Eye Surgery, New South Wales (Dr C Younan); Hawthorn Eye Clinic, Victoria (Dr L Chow); Les Manning, Queensland (Dr L Manning); Mark Perks Clinic, South Australia (Dr M Perks); Marsden Eye Specialists, New South Wales (Dr J Arnold, Dr H Cass); Melbourne Retina Associates, Victoria (Dr A Cohn); Mona Vale Eye Centre, New South Wales (Dr C Chung); Mosman Eye Centre, New South Wales (Dr C Chung); Nepean Valley Eye Surgeons, New South Wales (Dr G Banerjee); New England Eye Centre, New South Wales (Dr M Morgan); Port Macquarie Eye Centre, New South Wales (Dr J Game, Dr C Thompson); Retina & Macula Specialists (Hurstville), New South Wales (Dr R Chalasani, Dr M Chilov, Dr A Fung, Dr S Nothling); Retina & Macula Specialists (Miranda), New South Wales (Dr R Chalasani, Dr M Chilov, Dr S Nothling); Retina Associates, New South Wales (Dr S Fraser-Bell, Dr A Fung, Dr C Younan); Southern Eye Centre, Victoria (Dr D Louis); Specialist Eye Group, Victoria (Dr R Buttery, Dr L Chow, Dr A Cohn); St John of God Hospital Geelong, Victoria (Dr P Lockie); Strathfield Retina Clinic, New South Wales (Dr C Chung, Dr E Cornish); Sydney Eye Hospital, New South Wales (Dr S Fraser-Bell, Dr M Gillies); Tamworth Eye Centre, New South Wales (Dr P Hinchcliffe); Victoria Parade Eye Consultants, Victoria (Professor R Guymer); Victorian Eye Surgeons, Victoria (Dr A Cohn); Visionary Eye Specialists, New South Wales (Dr C Hooper).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors Research design: YH, ARH and MCG. Data acquisition and/or research execution: YH, ARH, JMW, GB, RF, RB, AF, JG, CYH, DB and MCG. Data analysis and/or interpretation: YH, ARH, CYH and MCG. Manuscript preparation: YH, ARH, JMW, GB, RF, RB, AF, JG, CYH, DB and MCG. YH is the guarantor.
Funding The Fight Retinal Blindness! Project is supported by a grant from the Macular Disease Foundation Australia and unrestricted educational grants from Bayer and Roche.
Map disclaimer The inclusion of any map (including the depiction of any boundaries therein), or of any geographic or locational reference, does not imply the expression of any opinion whatsoever on the part of BMJ concerning the legal status of any country, territory, jurisdiction or area or of its authorities. Any such expression remains solely that of the relevant source and is not endorsed by BMJ. Maps are provided without any warranty of any kind, either express or implied.
Competing interests MCG is a member of advisory boards for Novartis, Bayer, Apellis and Roche. DB is a consultant for Bayer and Alcon and received grants from Novartis (not ongoing). DB and MCG are inventors of the software used to collect the data for this analysis. ARH reports personal fees and others from Bayer and Roche.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.