[1]     INTRODUCTION

Diabetic retinopathy (DR) is a microvascular complication of diabetes mellitus and a leading cause of visual impairment and blindness worldwide. Its prevalence continues to increase each year, posing a significant public health challenge. The World Health Organization (WHO) estimates that the global prevalence of DR is approximately 22.97%, affecting around 103 million individuals, and is projected to reach 160 million by 2045.

DR accounts for approximately 4.8% of global blindness and represents one of the most common complications of diabetes after nephropathy. In Indonesia, the prevalence of DR is estimated at 43.1%, with 26.1% classified as vision-threatening cases. The economic burden associated with DR is substantial, including direct healthcare costs and indirect costs related to productivity loss and complications such as cardiovascular disease and depression.

Given the increasing burden of DR, effective strategies are required to prevent disease progression and reduce visual impairment. Early detection and timely treatment, including panretinal photocoagulation (PRP) and intravitreal anti–vascular endothelial growth factor (anti-VEGF) therapy, play a critical role in disease management.

Cost-effectiveness analysis (CEA) is an important tool for evaluating healthcare interventions by comparing costs and outcomes. It enables decision-makers to determine whether additional costs associated with a new intervention provide sufficient health benefits compared to existing treatment options. In this context, the incremental cost-effectiveness ratio (ICER) is used to quantify the additional cost per unit of health outcome gained.

This study aims to evaluate the cost-effectiveness of PRP alone compared with PRP combined with intravitreal anti-VEGF therapy for the management of diabetic retinopathy in a real-world hospital setting in Indonesia.

      [2]    MATERIAL & METHOD

This study was a retrospective cost-effectiveness analysis using secondary data obtained from medical records of patients with diabetic retinopathy treated at a private type B eye hospital in Surabaya, Indonesia, in 2021. This study was approved by the Institutional Ethics Committee of Universitas Padjadjaran with waiver of informed consent due to retrospective anonymized data use.

Study Perspective and Time Horizon

The economic evaluation was conducted from a societal perspective, incorporating both direct medical costs and indirect costs, including transportation expenses and productivity losses. The analysis adopted a one-year time horizon, reflecting the available follow-up period. Outcomes were evaluated at 6-month intervals, consistent with routine clinical follow-up for diabetic retinopathy. A one-year horizon was selected because complete follow-up data beyond 12 months were not consistently available in the retrospective dataset. Discounting was not applied because the analytic horizon was limited to one year. Although diabetic retinopathy is a chronic progressive disease, the selected timeframe reflects real-world treatment monitoring intervals and provides short-term evidence regarding economic impact in routine clinical practice.

Model Structure

A Markov model was developed to simulate disease progression and treatment outcomes. The model consisted of four mutually exclusive health states based on visual impairment severity:

(1) mild visual impairment,

(2) moderate visual impairment,

(3) severe visual impairment, and

(4) blindness.

All patients entered the model according to their baseline visual acuity and could transition between health states at each cycle. The model used a cycle length of 6 months, corresponding to clinical follow-up intervals. Blindness was treated as an absorbing state, with no transition to less severe states. Because baseline disease severity differed between treatment groups, subgroup interpretation was conducted cautiously. Formal adjustment methods such as propensity score matching were not feasible due to the retrospective aggregated nature of the dataset. Therefore, deterministic sensitivity analyses were performed to evaluate the robustness of results under varying assumptions regarding disease progression and treatment effectiveness.

Transition probabilities were estimated based on observed changes in visual acuity between follow-up visits and supplemented with evidence from published literature where necessary. The model assumed that treatment effectiveness was reflected in reduced progression to more severe health states. The Markov model assumed that all patients could transition between adjacent visual impairment states during each 6-month cycle according to treatment response and disease progression. Transition probabilities were also derived primarily from observed clinical changes in visual acuity during follow-up and supplemented using published evidence from diabetic retinopathy studies where local estimates were unavailable. Blindness was modeled as an absorbing state because recovery from complete visual loss is clinically uncommon in advanced diabetic retinopathy. The model further assumed that treatment effects remained constant throughout the one-year period and that adherence to treatment and follow-up schedules was maintained. Due to limitations in patient-level longitudinal data, transition probabilities were estimated using aggregated outcome distributions rather than individual trajectory analysis.

Costing Methodology

Costs were estimated in Indonesian Rupiah (IDR) for the price year 2021, ensuring consistency across all cost components. Both direct and indirect costs were included.

Direct medical costs comprised consultation fees, diagnostic examinations, procedural costs (including PRP and intravitreal anti-VEGF injections), medications, and hospital service charges. Unit costs for each component were obtained from hospital financial records and billing systems.

Indirect costs included transportation and productivity losses. Transportation costs were estimated based on commonly used modes of transport (motorcycle, car, and ride-hailing services), using average fuel consumption or fare estimates. Productivity losses were calculated using regional minimum wage data and the number of workdays lost due to treatment visits, assuming that time spent seeking care represents lost economic productivity. 

All costs were reported in IDR and will be converted to United States Dollars (USD) using the average exchange rate for 2021 (approximately IDR 14,300/USD) to facilitate international comparability as described on this Table 1.

Table 1. Per-Patient Cost Components of PRP Alone and PRP + Anti-VEGF Therapy (IDR, 2021)

Cost distributions between intervention groups reflected the hospital’s accounting and billing structure. In several cases, PRP-related services were recorded as bundled procedural packages, whereas anti-VEGF administration and operating room utilization were itemized separately in the combination therapy group. Furthermore, patients receiving PRP alone may have required more extensive laser treatment sessions, contributing to higher procedural expenditures. Therefore, differences in individual cost categories should be interpreted within the context of the hospital’s financial recording system rather than as isolated procedural comparison.

Cost and Outcome Measurement (DALYs)

Health outcomes were measured using Disability-Adjusted Life Years (DALYs), calculated as: DALY = YLL + YLD. In this study, Years of Life Lost (YLL) were assumed to be zero, as diabetic retinopathy is not a direct cause of mortality. This assumption is consistent with previous burden of disease studies in which DR contributes to morbidity rather than premature death. Therefore, DALYs were equivalent to Years Lived with Disability (YLD), calculated as: YLD = number of cases × disability weight. Disability weights were obtained from the Global Burden of Disease (GBD) 2019 study and categorized according to severity of visual impairment (moderate, severe, and blindness). The specific disability weights used are presented in Table 2. 

Table 2. Disability Weight

Visual acuity levels were mapped to corresponding health states in the Markov model, with each state assigned a specific disability weight. Improvements in visual acuity were modeled as transitions to less severe health states, resulting in lower disability weights and reduced YLD values. Averted cases were defined as the difference in the number of patients progressing to more severe visual impairment states between intervention groups over the study period. This reduction in disease progression directly contributed to lower DALYs in the intervention group.

Sensitivity Analysis

Probabilistic sensitivity analysis was not feasible because patient-level variance estimates and detailed transition uncertainty data were unavailable in the retrospective aggregated dataset. Consequently, deterministic one-way sensitivity analysis was used to evaluate parameter uncertainty. Future studies using individual-level longitudinal datasets should incorporate probabilistic modeling approaches to improve uncertainty estimation.

[3]   RESULTS

3.1 Respondent Characteristics

A total of 171 patients met the inclusion criteria and were included in the analysis. Baseline demographic and clinical characteristics are summarized in Table 3 to improve clarity and reduce redundancy.

The study population consisted of 86 males (50.29%) and 85 females (49.71%). The majority of patients were aged 48–57 years (39%) and 58–67 years (35%).

Differences in disease severity between intervention groups were observed. In the combination therapy group (PRP + anti-VEGF), patients were distributed across severity levels as follows: mild (30%), moderate (18%), severe (20%), and blindness (17%). In contrast, the PRP-only group consisted predominantly of mild (7%) and moderate (6%) cases, with fewer severe and blindness cases. Comparisons between groups were assessed using the Chi-square test for categorical variables. However, due to the use of aggregated retrospective data, these comparisons were considered exploratory and not intended for formal hypothesis testing.

3.2 Cost and Effectiveness Outcomes

The total cost of PRP combined with anti-VEGF therapy was IDR 16,771,000, compared to IDR 14,171,000 for PRP alone.

In terms of effectiveness, the combination therapy resulted in substantially greater health benefits, with 12.793 DALYs averted compared to 0.932 DALYs for PRP alone.

3.3 Incremental Cost-Effectiveness Analysis

The incremental cost associated with combination therapy was IDR 2,600,000, with an incremental effectiveness of 11.861 DALYs averted. This resulted in an ICER of IDR 219,206 per DALY averted.

3.4 Sensitivity Analysis

Due to the retrospective nature of the data and the use of aggregated estimates, formal statistical inference and confidence interval estimation were not feasible. However, deterministic one-way sensitivity analysis demonstrated that the ICER remained within the cost-effectiveness threshold across plausible variations in key parameters, indicating robustness of the findings.

3.5 Data Presentation

To address issues of repetition and improve clarity, all baseline characteristics and clinical distributions have been consolidated into structured tables. This approach ensures consistency and facilitates comparison between intervention groups.

[4]   DISCUSSION

This study demonstrates that PRP combined with intravitreal anti-VEGF therapy provides greater health benefits than PRP alone for diabetic retinopathy management in a real-world Indonesian hospital setting. Despite higher upfront treatment costs, the intervention remained cost-effective based on DALYs averted and established willingness-to-pay thresholds.

These findings are consistent with previous international studies evaluating anti-VEGF therapy in diabetic retinopathy and diabetic macular edema (DME). Clinical trials such as the RETAIN study have demonstrated that anti-VEGF regimens, including treat-and-extend approaches, achieve comparable or superior visual acuity outcomes while reducing treatment burden compared to conventional regimens. In addition, studies conducted in high-income countries have shown that anti-VEGF therapy is cost-effective due to its ability to prevent progression to severe visual impairment and blindness, thereby reducing long-term healthcare and societal costs. Although combination therapy demonstrated favorable cost-effectiveness, several implementation challenges should be considered within the Indonesian healthcare system. Access to retinal specialists and advanced ophthalmologic services remains concentrated in urban referral centers, limiting treatment availability in rural and remote areas. In addition, repeated anti-VEGF injections may impose logistical and financial burdens on patients despite national health insurance coverage through BPJS Kesehatan. Similar findings from other middle-income Asian countries, including Thailand and India, suggest that early retinal intervention may reduce long-term blindness-related productivity losses and healthcare expenditures. However, successful implementation requires strengthening diabetic retinopathy screening programs, improving referral pathways, and ensuring sustainable reimbursement mechanisms for retinal therapies. Given Indonesia’s increasing diabetes prevalence and growing economic burden of visual impairment, integrating cost-effective retinal interventions into national eye health strategies may provide substantial long-term societal benefit. Furthermore, economic evaluations in both high and middle income settings have reported favorable cost-effectiveness profiles for anti-VEGF therapy, particularly when targeted at patients with moderate to severe disease. Compared to previous studies in Indonesia, which were limited to short follow-up periods and focused only on direct medical costs, the present study provides a more comprehensive assessment by incorporating indirect costs and using DALYs as an outcome measure. This enhances the policy relevance of the findings for resource limited healthcare settings.

However, several sources of uncertainty should be considered when interpreting the results. First, the use of retrospective and aggregated data limits the ability to perform robust statistical inference and probabilistic sensitivity analysis. Second, the combination therapy group included a greater proportion of moderate-to-severe diabetic retinopathy cases at baseline, which may have underestimated the relative effectiveness of PRP + anti-VEGF therapy. Patients with more advanced disease generally have poorer visual prognosis and higher healthcare utilization. Therefore, despite this imbalance, the intervention remained cost-effective across sensitivity analyses, supporting the robustness of the findings, as the combination therapy group included a higher proportion of moderate-to-severe cases. Third, the Markov model relied on assumptions regarding disease progression and treatment effects, which may not fully capture real-world variability. The generalizability of these findings is also limited. The study was conducted in a single private eye hospital in Surabaya, and variations in clinical practice patterns, healthcare costs, and patient characteristics across regions in Indonesia may affect the applicability of the results. Therefore, caution is warranted when extrapolating these findings to the national level without further validation using multi-center data.

In Indonesia, implementation of anti-VEGF therapy faces several practical challenges, including limited retinal specialist availability, unequal access to ophthalmology services outside urban centers, and reimbursement constraints within BPJS Kesehatan. Despite these barriers, the increasing prevalence of diabetes mellitus and diabetic retinopathy highlights the importance of expanding access to early retinal screening and evidence-based treatment. Cost-effective interventions that prevent progression to blindness may reduce long-term societal burden and productivity loss, particularly among working-age populations.In Indonesia, implementation of anti-VEGF therapy faces several practical challenges, including limited retinal specialist availability, unequal access to ophthalmology services outside urban centers, and reimbursement constraints within BPJS Kesehatan. Despite these barriers, the increasing prevalence of diabetes mellitus and diabetic retinopathy highlights the importance of expanding access to early retinal screening and evidence-based treatment. Cost-effective interventions that prevent progression to blindness may reduce long-term societal burden and productivity loss, particularly among working-age populations.

Despite these limitations, the study has important implications for health policy and clinical practice. The ICER for combination therapy falls within the WHO-CHOICE threshold, indicating that it represents good value for money in the Indonesian context. These findings support the inclusion of anti-VEGF therapy in national health insurance schemes such as BPJS Kesehatan, particularly for patients at higher risk of disease progression. From a clinical perspective, integrating anti-VEGF therapy into standard treatment protocols may help reduce the long-term burden of visual impairment and associated economic consequences. The relatively short time horizon may underestimate the long-term benefits of anti-VEGF therapy, particularly the prevention of irreversible blindness and associated productivity losses. Longer modeling horizons would likely provide a more comprehensive estimation of lifetime economic value. 

Future research should incorporate multi-center data, longer time horizons, and probabilistic sensitivity analysis to improve the robustness and generalizability of findings. These findings support broader consideration of anti-VEGF therapy within diabetic retinopathy treatment strategies and national reimbursement systems. Nevertheless, further multi-center studies with longer follow-up periods and patient-level longitudinal data are needed to improve model precision and generalizability.

[5]   CONCLUSION

This study evaluated the cost-effectiveness of treatment strategies for diabetic retinopathy in a real-world hospital setting in Surabaya, Indonesia. Combination therapy demonstrated superior health outcomes with greater DALY averted despite higher initial treatment costs.

The incremental cost-effectiveness ratio (ICER) for the combination therapy was estimated at IDR 219,206 per DALY averted, which falls within the WHO-CHOICE cost-effectiveness threshold for Indonesia (2.86 times GDP per capita). These results suggest that combination therapy represents a cost-effective intervention for managing diabetic retinopathy.

From a policy perspective, the inclusion of anti-VEGF therapy alongside PRP may be considered in clinical practice and national health insurance schemes to reduce the burden of visual impairment. Further research using larger, multi-center datasets and longer time horizons is recommended to strengthen the evidence base.

[6] CONFLICT OF INTEREST:

The author declares no conflicts of interest

[7] FINANCIAL SUPPORT:

None.