Economic evaluation of topical ketorolac and nepafenac prophylaxis for early-stage retinopathy of prematurity: a multicenter real-world study in Colombia

Technical Summary: Economic Evaluation of Topical Ketorolac and Nepafenac Prophylaxis for Early-Stage ROP

Problem Statement
Retinopathy of prematurity (ROP) remains a leading cause of preventable childhood blindness, particularly in health systems where improvements in neonatal survival have not been matched by adequate screening capacity or specialist availability. In settings like Colombia, structural constraints—such as a limited workforce of ROP-trained specialists and significant geographic inequities in access—compromise timely surveillance and longitudinal follow-up. While early-stage ROP (Stage 1–2) often regresses spontaneously, a clinically significant proportion progresses to treatment-requiring disease, necessitating resource-intensive interventions like laser photocoagulation or anti-VEGF therapy.

In contexts where follow-up is uncertain, topical nonsteroidal anti-inflammatory drugs (NSAIDs), specifically ketorolac and nepafenac, have emerged as off-label prophylactic strategies to mitigate progression. However, the clinical evidence for these agents is limited and heterogeneous, and no prior study has evaluated the economic implications of such prophylaxis. This study addresses the gap in understanding whether early-stage pharmacologic prophylaxis represents a plausible use of constrained neonatal and ophthalmic resources.

Methodology
The study employed a multicenter retrospective cohort design to evaluate the real-world cost-effectiveness of topical ketorolac and nepafenac prophylaxis for Stage 1–2 ROP compared to a historical surveillance-only cohort.

• Setting and Population: The study included preterm infants screened for ROP across 14 facilities in Córdoba and Antioquia, Colombia, between January 2019 and September 2025. All infants were managed by a single retina specialist. The final analytic cohort comprised 192 infants: 48 receiving ketorolac, 98 receiving nepafenac, and 46 in the historical surveillance-only group (2011–2018).
• Economic Perspective: The evaluation was conducted from the health-system perspective, capturing direct medical costs attributable to ROP from the initial assessment through 60 weeks postmenstrual age. Costs included diagnostic evaluation, follow-up, prophylaxis, treatment of advanced disease, and retreatment. Unit costs were derived from the Colombian Ministry of Health tariff schedule and validated against local billing records.
• Effectiveness Metrics: Health-related quality of life was quantified using Quality-Adjusted Life-Years (QALYs) over the 60-week horizon. Utility weights were assigned to specific health states (e.g., early-stage ROP, regression, post-treatment states) based on structured expert judgment and literature, as validated preference-based instruments for premature neonates are unavailable.
• Analytical Approach: The study utilized probabilistic sensitivity analysis (PSA) with 100,000 Monte Carlo iterations. Costs were modeled using gamma distributions, and QALYs using beta distributions. The primary decision metric was Net Monetary Benefit (NMB), calculated against a willingness-to-pay (WTP) threshold of USD 5,180.80 per QALY (based on Colombian health-system standards). Incremental Cost-Effectiveness Ratios (ICERs) were also derived.
• Subgroup Analyses: Exploratory subgroup analyses were conducted based on gestational age (<28, 28–32, >32 weeks) and birth weight (<1,000 g, 1,000–1,500 g, >1,500 g).

Key Results

• Costs: Both prophylactic strategies were associated with lower mean ROP-attributable costs compared to surveillance alone. The mean cost for ketorolac was USD 430.30 (± 609.16) and for nepafenac was USD 436.60 (± 670.59), compared to USD 878.11 (± 961.49) for surveillance. However, these differences were not statistically significant (p = 0.5524 for ketorolac; p = 0.1591 for nepafenac).
• Effectiveness (QALYs): Ketorolac demonstrated a statistically significant increase in mean QALYs compared to surveillance (0.49 ± 0.06 vs. 0.46 ± 0.05; p = 0.0184). Nepafenac showed a higher mean QALY (0.47 ± 0.05) but the difference was not statistically significant (p = 0.2714).
• Subgroup Findings: The most consistent signal for QALY gains was observed in infants born at 28–32 weeks gestation, where both ketorolac (p = 0.0241) and nepafenac (p = 0.0057) showed significantly higher QALYs than surveillance.
• Probabilistic Analysis: Ketorolac demonstrated a higher Net Monetary Benefit (USD 598.18 ± 3.82) compared to nepafenac (USD 484.14 ± 3.84). The probability of ketorolac being cost-effective at the Colombian WTP threshold was 71.90%, compared to 68.16% for nepafenac. In the cost-effectiveness plane, 44.49% of ketorolac simulations fell in the dominant quadrant (lower cost, higher effectiveness), compared to 38.78% for nepafenac.

Key Contributions
This study provides the first economic evaluation of any pharmacologic prophylaxis for ROP. Its primary contributions include:

1. Real-World Evidence: It offers a multicenter, real-world assessment of off-label NSAID prophylaxis in a resource-constrained referral network, reflecting the challenges of variable follow-up and access to specialists.
2. Decision-Relevant Metrics: By integrating clinical trajectories with economic outcomes (QALYs and NMB), the study generates decision-relevant signals for policymakers and clinicians in settings where specialist scarcity is a critical barrier.
3. Agent-Specific Differentiation: The analysis distinguishes between the economic profiles of ketorolac and nepafenac, suggesting that while both may reduce costs, ketorolac presents a more favorable and consistent economic profile in this specific context.

Significance and Claims
The authors explicitly state that these results do not support the immediate routine adoption of topical NSAID prophylaxis for early-stage ROP. The study is modest in its claims, characterizing the findings as "hypothesis-generating" rather than definitive proof of cost-effectiveness.

The significance of the work lies in its identification of a "decision-relevant economic signal," particularly for ketorolac, which warrants further investigation. The authors conclude that the observed data justify the design and execution of well-powered, prospective randomized controlled trials to generate confirmatory evidence regarding safety, efficacy, and implementation feasibility before any broader clinical or policy implementation can be considered. The study serves as a rationale for structured evaluation rather than a mandate for immediate practice change.