Plans for assessment and collection of outcomes

All users of the study were trained and received a Certificate of Completion for Good Clinical practice. ETDRS BCVA will be determined after full refraction by masked, certified examiners using a certified room and certified equipment. The NEI VFQ-25 visual function questionnaire is a well-established, validated, patient-reported outcome measure (PROM) that has been used extensively in many trials. It is possible that loss of vision from ME has a particular impact on mental health. We therefore include the EQ5D5L questionnaire.

Plans to promote participant retention and complete follow-up

Visits, investigator meetings, and regular newsletters will all aim to impress on sites the importance of retention.

As patients in the experimental group, compared with the control group, undergo a laser treatment that is part of the routine care, we expect to obtain a satisfactory retention rate.

Study activity feedback will be periodically collected from recruited participants to identify any arising or recurring issues with the consent and recruitment, screening, laser procedure, or follow-up activities. Centers are requested to perform the examination of the trial within a dedicated consultation to minimize the time spent by patients and facilitate the follow-up. Participants who choose to discontinue will be asked to complete an early withdrawal visit (which will include the same datapoints as for the study month 12 exit visit), and the reason for withdrawal will be documented.

Data management

Source documents for clinical data are based on paper or electronic patient files depending on the organization of each center and specific questionnaires are collected on paper. Data entry for data made non-identifying are carried out on electronic media via a web browser. The study is conducted using the CleanWeb® electronic data capture system. The e-CRF is protected by a system of secure access (login and password). Data-entry are made directly via internet following different profiles and saved in an Oracle database. Data transmission to the web-server are performed by means of an Internet connection, and no specific software has to be installed at the study sites.

Anonymity of the subjects is guaranteed by using a patient ID number on all documents necessary for research. A dedicated data manager is employed at F-CRIN platform (AP-HP, Paris, France) to guide the entire data collection and to provide technical assistance. Periodically, this data manager analyzes the data base to identify centers for which the delay in data entry or the rate of missing values for some variables are unacceptable or absence of consistency between variables. In addition, clinical research assistants dedicated by the sponsor will monitor data accuracy in the different centers.

Confidentiality

During or after research involving human subjects, the data collected on the research subjects and sent to the sponsor by the investigators (or any other specialized parties) are made pseudonymous.

Under no circumstances shall the names and addresses of the subjects involved be displayed on the trial documents.

Only the surname and first name initials are recorded, accompanied by an encoded number specific to the study indicating the inclusion order of the subjects.

The sponsor will ensure that each research subject has given written permission for access to personal information about him or her which is strictly necessary for the quality control of the study. Data processing follows the European General Data Protection Regulation (GDPR) and the additional specific French rules (CNIL, the French Data Protection Authority).

Adverse event reporting and harms

All AEs (related or unrelated to the study), whether reported by the participant, discovered during questioning, directly observed, or detected by physical examination, laboratory test or other means, will be recorded in the participant’s medical record and on the appropriate study-specific case report forms (CRF), stating the duration and intensity of the event, action taken by the investigator, and outcome of the event. CTCAE will be used to collected and/or analyzed AEs.

As diabetic patients have many comorbidities and a high risk of systemic complications, unrelated to their participation to the present study, we decided to exclude from the AEs the following situations:

Depending on the diabetes stage, patients may be hospitalized in a day-in unit or in the ward for a few days for the annual screening, for an acute glucose deterioration, or for the progression or for any diabetes complication (retinal detachment, coronary heart disease, acute kidney failure…).

Thus, normal or natural evolution of the study may lead to:

Hospitalization scheduled for routine screening.

Hospitalization for progression of a complication targeting the cardiovascular system, kidney function, painful neuropathy, foot ulcer.

#Special circumstances that do not need notification from investigators:

Hospitalization for a pre-existing condition/disease.

Hospitalization for medical or surgical treatment that has been scheduled before the study inclusion.

Hospitalization for social or administrative reason.

Hospitalization through the emergency room for diabetes or diabetes related-complications.

As regards laser therapy, retinal scars, increase of such scars with time, potentially involving foveal zone, and secondary choroidal neovascularization are the main expected complications. They will be studied using autofluorescence imaging from angiography frames at M3 M6 and M12 and using OCT B-scan study of external retina. In case of occurrence, they will be reported in the “comment item” located in the “imaging section” of the eCRF. Given prior experience of anti-VEGF therapy, the major adverse events expected are as follows: endophthalmitis, traumatic cataract related to IVI, retinal detachment, and potential increased risk of systemic adverse event such as acute myocardial infarction, cardiovascular disease, or kidney disease [35]. Those events will be collected and reported as SAE. The initial report, the SAE follow-up reports, and all other documents must be sent to the sponsor (by delegation the pharmacovigilance-CRO) by email to safety@fordrugconsulting.fr.

Frequency and plans for auditing trial conduct

To ensure the safety and respect of those individuals who have agreed to participate in the study, the sponsor will implement a quality assurance system to best monitor the running of the study in the investigation centers.

The medico-economic evaluation is of outmost importance given the economic burden of macular edema. Therefore, a health economics analysis will be performed as part of the study. Its objective will be to evaluate a strategy based on IGTL associated with anti-VEGF compared with anti-VEGF alone and to estimate the incremental cost per incremental quality adjusted life year (QALY). In non-inferiority studies, the cost-effectiveness analysis allows an appropriate representation of uncertainty, rather than hypothesis testing. We will represent the distribution of the joint density of mean cost and effect differences by using bootstrap replications. This removes the focus on hypothesis testing which leads to an overemphasis on type I errors and allows guidelines developers and policy makers to set their own thresholds for the probability that an intervention is acceptable [36]. This method of obtaining a confidence interval for cost-effectiveness from the cost-effectiveness plane and the acceptability curve does not give a confidence interval on the incremental cost-effectiveness ratio (ICER) statistic, as the ceiling ratio is defined only in positive quadrants of the cost-effectiveness plane. The statistical problems associated with negative ICERs (for example if the added targeted ICG laser therapy treatment is not superior but cheaper with reduced use of anti-VEGF) are avoided. Also, the results of the economic evaluation allow policy makers to set their own confidence level (not bound by the 5% alpha) and cost-effectiveness threshold.

The evaluation follows the recommendations from the French national health authority and the reporting will follow the CHEERS statement (https://www.has-sante.fr/jcms/r_1499251/fr/choix-methodologiques-pour-l-evaluation-economique-a-la-has/https://www.equator-network.org/reporting-guidelines/cheers). The perspective chosen is the healthcare system/payer and the patients, and the time horizon is 1 year. This choice is justified by the fact that studies in other countries have chosen the payers’ viewpoint; French patients with diabetic macular edema are eligible for 100% coverage of healthcare spending. Some patients however may elect to consult self-employed ophthalmologists during the study period and incur out of pocket costs for extra billing, which will not be captured. The economic evaluation is based on the entire population of patients included in the trial. Resources are collected prospectively at the patient level. The study is planned, undertaken, and analyzed according to the intention-to-treat principle. The unit of analysis is the patient. Because of the short duration of the follow-up, no discounting is required.

The costs of the laser will be estimated from microcosting for the procedure itself, collecting data on equipment and consumable use, staff time, and expected patient volume and usual rules for depreciation costs.

Resource utilization (we expect that patients will be managed as outpatients predominantly and that no hospital admissions will occur in relation to the protocol) will be collected at the patient level, partly via the study CRF (use of anti-VEGF and other treatment related to the edema) and partly via patient questioning during the follow-up visits. We will also ask patients about healthcare utilization for their eyes outside the study protocol and possible out of pocket payments.

Resource utilization related to adverse events will be systematically recorded.

All resources will be valued using the current list prices (drugs) and tariffs (consultations and tests). Hospital admissions will be valued using the current DGR costs.

The quality of life will be assessed in both groups at baseline, 3 months, 6 months, and at the end of the study. EQ 5D 5L scores will be valued using French tariffs [37]. The utility values will be attributed to the time period corresponding to mid-point between data collection. The primary outcome measure is cumulated QALYs. The utilities will be converted into QALYs for each arm using the area under the curve (AUC) method. This method assumes a linear relationship between values at different time points [38]. The gain of QALYs will be the difference between QALYs calculated in each arm. We will also compute the differential adjusted QALYs using regression models using the utility at baseline as an independent variable. The cost-effectiveness of IGTL combined to anti-VEGF vs anti-VEGF alone will be the incremental cost divided by incremental QALYs. The utility values will be attributed to the time period corresponding to mid-point between data collection. The difference in QALYs will be estimated as the difference in the area between the utility curves for the two treatment groups.

The economic endpoint is expressed as the point estimate of the incremental cost-effectiveness ratio (ICERs): where Δ costs (between groups)/Δ QALYs (between groups). In the case of non-inferiority studies, the innovation treatment is potentially decrementally cost-effective. In other words, the performance could be acceptably lower than reference strategy but result in a lower overall cost. If such a result is reproduced here, we plan subgroup analyses to identify which patients would be affected by a decrease in performance. The result is compared with the accepted French threshold values [39].

Baseline results will be presented as mean ± SD, median interquartile ranges (IQR), or as frequencies with percentages. Resource use data will be presented as means with standard error of the mean despite non-normal distribution because they better represent per patient data than median values and compared using nonparametric testing. Costs and QALYs will be presented as means with 2.5 to 97.5% bootstrapped intervals. Between-group comparisons of costs will be performed using the bootstrap t-test. Between-group comparisons of QALYs will be performed using nonparametric testing. A joint comparison of costs and QALYs will be performed by nonparametric bootstrapping with 1000 resamples. The uncertainty surrounding the ICER will be presented on the cost-effectiveness plane and acceptability curves.

All costs will be reported in € at the end of the study.