Implementing multi-level network meta-regression for time-to-event outcomes: A case study in relapsed refractory multiple myeloma.

OBJECTIVES: Multi-level network meta-regression (ML-NMR) leverages individual patient data (IPD) and aggregate data (AD) from a network of randomized controlled trials (RCTs) to assess the comparative efficacy of multiple treatments, while adjusting for between-study differences. We provide an overview of ML-NMR for time-to-event outcomes and apply it to an illustrative case study, including example R code. METHODS: The case study evaluated the comparative efficacy of idecabtagene vicleucel (ide-cel), selinexor+dexamethasone (Sd), belantamab mafodotin (BM), and conventional care (CC) for patients with triple class exposed relapsed/refractory multiple myeloma in terms of overall survival (OS). Single-arm clinical trials and real-world data were naively combined to create an AD artificial RCT (aRCT) (MAMMOTH-CC versus DREAMM-2-BM verus STORM-2-Sd) and an IPD aRCT (KarMMa-ide-cel versus KarMMa-RW-CC). With some assumptions, we incorporated continuous covariates with skewed distributions, reported as median and range. The ML-NMR models adjusted for number of prior lines, triple-class refractory (TCR) status, and age and were compared via the leave-one-out information criterion (LOOIC). We summarized predicted hazard ratios and survival (95% credible intervals) in the IPD aRCT population. RESULTS: The Weibull ML-NMR model had the lowest LOOIC. Ide-cel was more efficacious than Sd, BM, and CC in terms of OS. Effect modifiers had minimal impact on the model and only TCR was a prognostic factor. CONCLUSIONS: We demonstrate an application of ML-NMR for time-to-event outcomes and introduce code that can be used to aid implementation. Given its benefits, we encourage practitioners to utilize ML-NMR when population adjustment is necessary for comparisons of multiple treatments.

Challenges of Incorporating Life Cycle Drug Pricing in Cost-Effectiveness Models: A Review of Methods and Modeling Suggestions.

OBJECTIVES: This study aimed to conduct a review of existing methods used to incorporate life cycle drug pricing (LCDP) in cost-effectiveness analyses (CEAs), identify common methodological challenges, and suggest modeling approaches for prospectively implementing LCDP in CEA. METHODS: Two complementary searches were conducted in PubMed, combined with hand searching and reference mining, to identify English language full-text articles that explored (1) how drug prices change over time and (2) methods used to apply dynamic pricing in cost-effectiveness models (CEMs). Relevant articles were reviewed, and authors discussed the common methodological practices used in the literature and their associated challenges on prospectively implementing LCDP in CEMs. For each key challenge identified, we provide modeling suggestions to address the issue. RESULTS: We screened 1200 studies based on title and abstract; 117 were reviewed for eligibility, and 47 individual studies were included across both searches. Variations in prices over a product's life cycle are complex and multifactorial, and models applying LCDP in CEA varied in their methodology. We identified 4 key challenges to modeling LCDP in CEA, including how to model price trends before and after loss of exclusivity, how to capture the effect of price changes on future patient cohorts, and how to report results. CONCLUSION: Accurately quantifying the impact of LCDP requires careful consideration of multiple aspects pertaining to both the evolution of drug prices and how to reflect these in CEA. Although uncertainties remain, our findings can aid implementation and evaluation of LCDP in economic evaluations.

Artificial Intelligence to Automate Health Economic Modelling: A Case Study to Evaluate the Potential Application of Large Language Models.

BACKGROUND: Current generation large language models (LLMs) such as Generative Pre-Trained Transformer 4 (GPT-4) have achieved human-level performance on many tasks including the generation of computer code based on textual input. This study aimed to assess whether GPT-4 could be used to automatically programme two published health economic analyses. METHODS: The two analyses were partitioned survival models evaluating interventions in non-small cell lung cancer (NSCLC) and renal cell carcinoma (RCC). We developed prompts which instructed GPT-4 to programme the NSCLC and RCC models in R, and which provided descriptions of each model's methods, assumptions and parameter values. The results of the generated scripts were compared to the published values from the original, human-programmed models. The models were replicated 15 times to capture variability in GPT-4's output. RESULTS: GPT-4 fully replicated the NSCLC model with high accuracy: 100% (15/15) of the artificial intelligence (AI)-generated NSCLC models were error-free or contained a single minor error, and 93% (14/15) were completely error-free. GPT-4 closely replicated the RCC model, although human intervention was required to simplify an element of the model design (one of the model's fifteen input calculations) because it used too many sequential steps to be implemented in a single prompt. With this simplification, 87% (13/15) of the AI-generated RCC models were error-free or contained a single minor error, and 60% (9/15) were completely error-free. Error-free model scripts replicated the published incremental cost-effectiveness ratios to within 1%. CONCLUSION: This study provides a promising indication that GPT-4 can have practical applications in the automation of health economic model construction. Potential benefits include accelerated model development timelines and reduced costs of development. Further research is necessary to explore the generalisability of LLM-based automation across a larger sample of models.

Artificial Intelligence to Automate Network Meta-Analyses: Four Case Studies to Evaluate the Potential Application of Large Language Models.

BACKGROUND: The emergence of artificial intelligence, capable of human-level performance on some tasks, presents an opportunity to revolutionise development of systematic reviews and network meta-analyses (NMAs). In this pilot study, we aim to assess use of a large-language model (LLM, Generative Pre-trained Transformer 4 [GPT-4]) to automatically extract data from publications, write an R script to conduct an NMA and interpret the results. METHODS: We considered four case studies involving binary and time-to-event outcomes in two disease areas, for which an NMA had previously been conducted manually. For each case study, a Python script was developed that communicated with the LLM via application programming interface (API) calls. The LLM was prompted to extract relevant data from publications, to create an R script to be used to run the NMA and then to produce a small report describing the analysis. RESULTS: The LLM had a > 99% success rate of accurately extracting data across 20 runs for each case study and could generate R scripts that could be run end-to-end without human input. It also produced good quality reports describing the disease area, analysis conducted, results obtained and a correct interpretation of the results. CONCLUSIONS: This study provides a promising indication of the feasibility of using current generation LLMs to automate data extraction, code generation and NMA result interpretation, which could result in significant time savings and reduce human error. This is provided that routine technical checks are performed, as recommend for human-conducted analyses. Whilst not currently 100% consistent, LLMs are likely to improve with time.

The Clinical Emergency Data Registry: Structure, Use, and Limitations for Research.

The Clinical Emergency Data Registry (CEDR) is a qualified clinical data registry that collects data from participating emergency departments (EDs) in the United States for quality measurement, improvement, and reporting purposes. This article aims to provide an overview of the data collection and validation process, describe the existing data structure and elements, and explain the potential opportunities and limitations for ongoing and future research use. CEDR data are primarily collected for quality reporting purposes and are obtained from diverse sources, including electronic health records and billing data that are de-identified and stored in a secure, centralized database. The CEDR data structure is organized around clinical episodes, which contain multiple data elements that are standardized using common data elements and are mapped to established terminologies to enable interoperability and data sharing. The data elements include patient demographics, clinical characteristics, diagnostic and treatment procedures, and outcomes. Key limitations include the limited generalizability due to the selective nature of participating EDs and the limited validation and completeness of data elements not currently used for quality reporting purposes, including demographic data. Nonetheless, CEDR holds great potential for ongoing and future research in emergency medicine due to its large-volume, longitudinal, near real-time, clinical data. In 2021, the American College of Emergency Physicians authorized the transition from CEDR to the Emergency Medicine Data Institute, which will catalyze investments in improved data quality and completeness for research to advance emergency care.

Retrospective Comparison of Survival Projections for CAR T-Cell Therapies in Large B-Cell Lymphoma.

BACKGROUND: Durable remission has been observed in patients with relapsed or refractory (R/R) large B-cell lymphoma (LBCL) treated with chimeric antigen receptor (CAR) T-cell therapy. Consequently, hazard functions for overall survival (OS) are often complex, requiring the use of flexible methods for extrapolations. OBJECTIVES: We aimed to retrospectively compare the predictive accuracy of different survival extrapolation methods and evaluate the validity of goodness-of-fit (GOF) criteria-based model selection for CAR T-cell therapies in R/R LBCL. METHODS: OS data were sourced from JULIET, ZUMA-1, and TRANSCEND NHL 001. Standard parametric, mixture cure, cubic spline, and mixture models were fit to multiple database locks (DBLs), with varying follow-up durations. GOF was assessed using the Akaike information criterion and Bayesian information criterion. Predictive accuracy was calculated as the mean absolute error (MAE) relative to OS observed in the most mature DBL. RESULTS: For all studies, mixture cure and cubic spline models provided the best predictive accuracy for the least mature DBL (MAE 0.013‒0.085 and 0.014‒0.128, respectively). The predictive accuracy of the standard parametric and mixture models showed larger variation (MAE 0.024‒0.162 and 0.013‒0.176, respectively). With increasing data maturity, the predictive accuracy of standard parametric models remained poor. Correlation between GOF criteria and predictive accuracy was low, particularly for the least mature DBL. CONCLUSIONS: Our analyses demonstrated that mixture cure and cubic spline models provide the most accurate survival extrapolations of CAR T-cell therapies in LBCL. Furthermore, GOF should not be the only criteria used when selecting the optimal survival model.

Challenges in conducting fractional polynomial and standard parametric network meta-analyses of immune checkpoint inhibitors for first-line advanced renal cell carcinoma.

Aim: Network meta-analyses (NMAs) increasingly feature time-varying hazards to account for non-proportional hazards between different drug classes. This paper outlines an algorithm for selecting clinically plausible fractional polynomial NMA models. Methods: The NMA of four immune checkpoint inhibitors (ICIs) + tyrosine kinase inhibitors (TKIs) and one TKI therapy for renal cell carcinoma (RCC) served as case study. Overall survival (OS) and progression free survival (PFS) data were reconstructed from the literature, 46 models were fitted. The algorithm entailed a-priori face validity criteria for survival and hazards, based on clinical expert input, and predictive accuracy against trial data. Selected models were compared with statistically best-fitting models. Results: Three valid PFS and two OS models were identified. All models overestimated PFS, the OS model featured crossing ICI + TKI versus TKI curves as per expert opinion. Conventionally selected models showed implausible survival. Conclusion: The selection algorithm considering face validity, predictive accuracy, and expert opinion improved the clinical plausibility of first-line RCC survival models.

A review of HTA guidelines on societal and novel value elements.

OBJECTIVES: Health technology assessment (HTA) organizations vary in terms of how they conduct assessments. We assess whether and to what extent HTA bodies have adopted societal and novel elements of value in their economic evaluations. METHODS: After categorizing "societal" and "novel" elements of value, we reviewed fifty-three HTA guidelines. We collected data on whether each guideline mentioned each societal or novel element of value, and if so, whether the guideline recommended the element's inclusion in the base case, sensitivity analysis, or qualitative discussion in the HTA. RESULTS: The HTA guidelines mention on average 5.9 of the twenty-one societal and novel value elements we identified (range 0-16), including 2.3 of the ten societal elements and 3.3 of the eleven novel value elements. Only four value elements (productivity, family spillover, equity, and transportation) appear in over half of the HTA guidelines, whereas thirteen value elements are mentioned in fewer than one-sixth of the guidelines, and two elements receive no mention. Most guidelines do not recommend value element inclusion in the base case, sensitivity analysis, or qualitative discussion in the HTA. CONCLUSIONS: Ideally, more HTA organizations will adopt guidelines for measuring societal and novel value elements, including analytic considerations. Importantly, simply recommending in guidelines that HTA bodies consider novel elements may not lead to their incorporation into assessments or ultimate decision making.

Next-generation enhanced-efficiency fertilizers for sustained food security.

Nitrogen losses in agricultural systems can be reduced through enhanced-efficiency fertilizers (EEFs), which control the physicochemical release from fertilizers and biological nitrogen transformations in soils. The adoption of EEFs by farmers requires evidence of consistent performance across soils, crops and climates, paired with information on the economic advantages. Here we show that the benefits of EEFs due to avoided social costs of nitrogen pollution considerably outweigh their costs-and must be incorporated in fertilizer policies. We outline new approaches to the design of EEFs using enzyme inhibitors with modifiable chemical structures and engineered, biodegradable coatings that respond to plant rhizosphere signalling molecules.

The Potential of Deep Roots to Mitigate Impacts of Heatwaves and Declining Rainfall on Pastures in Southeast Australia.

Declines in growing-season rainfall and increases in the frequency of heatwaves in southern Australia necessitate effective adaptation. The Sustainable Grazing Systems Pasture Model (SGS) was used to model the growth of three pasture species differing in root depth and root distribution under three different climate scenarios at two sites. The modelled metabolisable energy intake (in MJ) was used in a partial discounted net cash flow budget. Both the biophysical and economic modelling suggest that deep roots were advantageous in all climate scenarios at the long growing season site but provided no to little advantage at the short growing season site, likely due to the deep-rooted species drying out the soil profile earlier. In scenarios including climate change, the DM production of the deep-rooted species at the long growing season site averaged 386 kg/ha/year more than the more shallow-rooted species, while at the site with a shorter growing season it averaged 205 kg/ha/year less than the shallower-rooted species. The timing of the extra growth and pasture persistence strongly influenced the extent of the benefit. At the short growing season site other adaptation options such as summer dormancy will likely be necessary.

Economic Analysis of Offering Different Herbage Allowances to Dairy Cows Fed a Partial Mixed Ration.

The economics of grazing dairy cows offered a range of herbage allowances and fed supplements as a partial mixed ration (PMR) were examined where profit was defined as the margin between total milk income and the cost of pasture plus PMR supplement. The analysis made use of milk production and feed intake data from two dairy cow nutrition experiments, one in early lactation and the other in late lactation. In early lactation and at a PMR intake of 6 kg DM/cow per day, the profit from the cows with access to a medium herbage allowance (25 kg DM/cow per day) was AUD 1.40/cow per day higher than that for cows on a low allowance (15 kg DM/cow per day). At a higher PMR intake of 14 kg DM/cow per day, the profit from the cows on a medium herbage allowance was AUD 0.45/cow per day higher than the cows on a low allowance; there was no additional profit from increasing the herbage allowance from medium to high (40 kg DM/cow per day). In late lactation, the profit from the cows fed a PMR with a medium herbage allowance (20 kg DM/cow per day) was only higher than the cows on a low allowance (12 kg DM/cow per day) when the PMR intake was between 6 and 12 kg DM/cow per day. There was also a difference of AUD +0.50/cow per day between the PMR with medium and high herbage allowance (32 kg DM/cow per day). It was concluded that farmers who feed a PMR to dairy cows should offer at least a medium herbage allowance to optimize profit. While feeding additional PMR increases milk production and profit, further gains would be available by offering a higher herbage allowance. These findings provide an estimate of the net benefits of different herbage allowances when feeding a PMR and will enable farmers to manage their feeding systems more profitably.

What Did Time Tell Us? A Comparison and Retrospective Validation of Different Survival Extrapolation Methods for Immuno-Oncologic Therapy in Advanced or Metastatic Renal Cell Carcinoma.

BACKGROUND: The immuno-oncologic (IO) mechanism of action may lead to an overall survival (OS) hazard that changes over time, producing shapes that standard parametric extrapolation methods may struggle to reflect. Furthermore, selection of the most appropriate extrapolation method for health technology assessment is often based on trial data with limited follow-up. OBJECTIVE: To examine this problem, we fitted a range of extrapolation methods to patient-level survival data from CheckMate 025 (NCT01668784, CM-025), a phase III trial comparing nivolumab with everolimus for previously treated advanced renal cell carcinoma (aRCC), to assess their predictive accuracy over time. METHODS: Six extrapolation methods were examined: standard parametric models, natural cubic splines, piecewise models combining Kaplan-Meier data with an exponential or non-exponential distribution, response-based landmark models, and parametric mixture models. We produced three database locks (DBLs) at minimum follow-ups of 15, 27, and 39 months to align with previously published CM-025 data. A three-step evaluation process was adopted: (1) selection of the distribution family for each method in each of the three DBLs, (2) internal validation comparing extrapolation-based landmark and mean survival with the latest CM-025 dataset (minimum follow-up, 64 months), and (3) external validation of survival projections using clinical expert opinion and long-term follow-up data from other nivolumab studies in aRCC (CheckMate 003 and CheckMate 010). RESULTS: All extrapolation methods, with the exception of mixture models, underestimated landmark and mean OS for nivolumab compared with CM-025 long-term follow-up data. OS estimates for everolimus tended to be more accurate, with four of the six methods providing landmark OS estimates within the 95% confidence interval of observed OS as per the latest dataset. The predictive accuracy of survival extrapolation methods fitted to nivolumab also showed greater variation than for everolimus. The proportional hazards assumption held for all DBLs, and a dependent log-logistic model provided reliable estimates of longer-term survival for both nivolumab and everolimus across the DBLs. Although mixture models and response-based landmark models provided reasonable estimates of OS based on the 39-month DBL, this was not the case for the two earlier DBLs. The piecewise exponential models consistently underestimated OS for both nivolumab and everolimus at clinically meaningful pre-specified landmark time points. CONCLUSIONS: This aRCC case study identified marked differences in the predictive accuracy of survival extrapolation methods for nivolumab but less so for everolimus. The dependent log-logistic model did not suffer from overfitting to early DBLs to the same extent as more complex methods. Methods that provide more degrees of freedom may accurately represent survival for IO therapy, particularly if data are more mature or external data are available to inform the long-term extrapolations.

Economic Threshold Analysis of Supplementing Dairy Cow Diets with Betaine and Fat during a Heat Challenge: A Pre- and Post-Experimental Comparison.

Ex ante economic analysis can be used to establish the production threshold for a proposed experimental diet to be as profitable as the control treatment. This study reports (1) a pre-experimental economic analysis to estimate the milk production thresholds for an experiment where dietary supplements were fed to dairy cows experiencing a heat challenge, and (2) comparison of these thresholds to the milk production results of the subsequent animal experiment. The pre-experimental thresholds equated to a 1% increase in milk production for the betaine supplement, 9% increase for the fat supplement, and 11% increase for fat and betaine in combination, to achieve the same contribution to farm profit as the control diet. For the post-experimental comparison, previously modelled climate predictions were used to extrapolate the milk production results from the animal experiment over the annual hot-weather period for the dairying region in northern Victoria, Australia. Supplementing diets with fat or betaine had the potential to produce enough extra milk to exceed the production thresholds, making either supplement a profitable alternative to feeding the control diet during the hot-weather period. Feeding fat and betaine in combination failed to result in the extra milk required to justify the additional cost when compared to the control diet.

Comparison of accelerometer-derived physical activity levels between individuals with and without cancer: a UK Biobank study.

Aim: To identify the difference in physical activity (PA) levels between individuals with and without cancer, and to estimate all-cause mortality associated with this difference. Methods: Current cancer, cancer survivor and cancer-free groups were identified from the UK Biobank. We used multivariate and Cox regression to estimate PA differences and association of PA with all-cause mortality. Results: Compared with the cancer-free individuals, participants in the two cancer groups had fewer minutes in moderate-to-vigorous PA per day in adjusted analyses. The PA difference was associated with higher mortality in the current cancer group. Conclusion: Patients with a history of cancer were less active than those without cancer, and PA is associated with increased mortality. PA improvement strategies in cancer patients must be explored.

Automated extraction of treatment patterns from social media posts: an exploratory analysis in renal cell carcinoma.

Aim: The use of health-related social media forums by patients is increasing and the size of these forums creates a rich record of patient opinions and experiences, including treatment histories. This study aimed to understand the possibility of extracting treatment patterns in an automated manner for patients with renal cell carcinoma, using natural language processing, rule-based decisions, and machine learning. Patients & methods: Obtained results were compared with those from published observational studies. Results: 42 comparisons across seven therapies, three lines of treatment, and two-time periods were made; 37 of the social media estimates fell within the variation seen across the published studies. Conclusion: This exploratory work shows that estimating treatment patterns from social media is possible and generates results within the variation seen in published studies, although further development and validation of the approach is needed.

Real-world data and the patient perspective: the PROmise of social media?

Understanding the patient perspective is fundamental to delivering patient-centred care. In most healthcare systems, however, patient-reported outcomes are not regularly collected or recorded as part of routine clinical care, despite evidence that doing so can have tangible clinical benefit. In the absence of the routine collection of these data, research is beginning to turn to social media as a novel means to capture the patient voice. Publicly available social media data can now be analysed with relative ease, bypassing many logistical hurdles associated with traditional approaches and allowing for accelerated and cost-effective data collection. Existing work has shown these data can offer credible insight into the patient experience, although more work is needed to understand limitations with respect to patient representativeness and nuances of captured experience. Nevertheless, linking social media to electronic medical records offers a significant opportunity for patient views to be systematically collected for health services research and ultimately to improve patient care.

Validity of social media for assessing treatment patterns in oncology patients: a case study in melanoma.

There is a need to understand how patients are managed in the real world to better understand disease burden and unmet need. Traditional approaches to gather these data include the use of electronic medical record (EMR) or claims databases; however, in many cases data access policies prevent rapid insight gathering. Social media may provide a potential source of real-world data to assess treatment patterns, but the limitations and biases of doing so have not yet been evaluated. Here, we assessed whether patient treatment patterns extracted from publicly available patient forums compare to results from more traditional EMR and claims databases. We observed that the 95% confidence intervals of proportions of treatments received at first, second, and third line for advanced/metastatic melanoma generated from unstructured social media data overlapped with 95% confidence intervals from proportions obtained from 1 or more traditional EMR/Claims databases. Social media may offer a valid data option to understand treatment patterns in the real world.

Application of dynamic modeling for survival estimation in advanced renal cell carcinoma.

OBJECTIVE: In oncology, extrapolation of clinical outcomes beyond trial duration is traditionally achieved by parametric survival analysis using population-level outcomes. This approach may not fully capture the benefit/risk profile of immunotherapies due to their unique mechanisms of action. We evaluated an alternative approach-dynamic modeling-to predict outcomes in patients with advanced renal cell carcinoma. We compared standard parametric fitting and dynamic modeling for survival estimation of nivolumab and everolimus using data from the phase III CheckMate 025 study. METHODS: We developed two statistical approaches to predict longer-term outcomes (progression, treatment discontinuation, and survival) for nivolumab and everolimus, then compared these predictions against follow-up clinical trial data to assess their proximity to observed outcomes. For the parametric survival analyses, we selected a probability distribution based on its fit to observed population-level outcomes at 14-month minimum follow-up and used it to predict longer-term outcomes. For dynamic modeling, we used a multivariate Cox regression based on patient-level data, which included risk scores, and probability and duration of response as predictors of longer-term outcomes. Both sets of predictions were compared against trial data with 26- and 38-month minimum follow-up. RESULTS: Both statistical approaches led to comparable fits to observed trial data for median progression, discontinuation, and survival. However, beyond the trial duration, mean survival predictions differed substantially between methods for nivolumab (30.8 and 51.5 months), but not everolimus (27.2 and 29.8 months). Longer-term follow-up data from CheckMate 025 and phase I/II studies resembled dynamic model predictions for nivolumab. CONCLUSIONS: Dynamic modeling can be a good alternative to parametric survival fitting for immunotherapies because it may help better capture the longer-term benefit/risk profile and support health-economic evaluations of immunotherapies.

Agribusiness Perspectives on Transgenic Wheat.

Declining yields of the major human food crops, looming growth in global population and rise of populism, and ill-founded bans on agricultural and horticultural crops and foodstuffs which are genetically modified have potentially serious implications. It makes the chance less than otherwise would be the case that agribusiness value chains in the future will meet the growing demand around the world for more and different foods from more and wealthier people. In the agribusiness value chain, transgenic wheat, meeting a consumer "trigger need" also must meet the "experience" and "credence," risk-related criteria of well-informed consumers. Public policy that rejects science-based evidence about the reductions in costs of production and price of genetically modified agricultural products and the science about the safety of genetically modified foods, including transgenic wheat, has imposed significant costs on producers and consumers. If the science-based evidence is accepted, transgenic wheat has potential to improve significantly the well-being of grain growers and consumers all over the world.

Simulation Modelling in Ophthalmology: Application to Cost Effectiveness of Ranibizumab and Aflibercept for the Treatment of Wet Age-Related Macular Degeneration in the United Kingdom.

BACKGROUND: Previously developed models in ophthalmology have generally used a Markovian structure. There are a number of limitations with this approach, most notably the ability to base patient outcomes on best-corrected visual acuity (BCVA) in both eyes, which may be overcome using a different modelling structure. Simulation modelling allows for this to be modelled more precisely, and therefore may provide more accurate and relevant estimates of the cost effectiveness of ophthalmology interventions. OBJECTIVE: This study aimed to explore the appropriateness of simulation modelling in ophthalmology, using the disease area of wet age-related macular degeneration (wAMD) as an example. METHODS: A de novo economic model was built using a patient-level simulation, which compared ranibizumab with aflibercept in wAMD. Disease progression was measured using BCVA. Health-related quality of life (HRQoL) was estimated using a regression analysis linking BCVA in each eye to utility. The analysis was from the perspective of the National Health Service in the UK. Five different regression models were explored and were based on BCVA in either one eye or both eyes. RESULTS: The model outputs provide some evidence to support the hypothesis that the analyses using the two-eye models for estimating HRQoL generate a more accurate estimation of incremental quality-adjusted life-years (QALYs) associated with the positive treatment effect for ranibizumab versus aflibercept. Second-order analysis broadly supported these findings, and showed that the variation in incremental costs was slightly lower than in incremental QALYs. The second-order analysis estimated similar incremental costs and a greater overall variation in incremental QALYs than the first-order analysis, suggesting important non-linearities within the model. CONCLUSIONS: This analysis suggests that patient-level simulation models may be well suited to representing the real-world patient pathway in wAMD, particularly when aspects of disease progression cannot be adequately captured using a Markov structure. The benefits of a simulation approach can be demonstrated in the modelling of HRQoL as a function of BCVA in both eyes.