Why did IVI choose MDD as its next open-source economic model?
IVI chose MDD as the disease state for the third OSVP model based on its prevalence, significant societal burden, and broad interest among multiple stakeholder groups (e.g., patients, employers) to seek better treatments and outcomes in an efficient manner. IVI also sought to develop a model that assesses both pharmaceutical and non-pharmaceutical treatment options to inform a broader discussion about value.
What is the objective of the MDD Model?
The objective in developing a model through the OSVP is not to produce a single assessment for a single treatment option with a specific set of assumptions and estimates. Rather, it is to explore and test ways to improve how we develop economic models to assess value, improve alignment with real-world decision needs, and advance dialogue about how best to use various methods (e.g., economic modeling) to inform resource allocation and other decisions in healthcare.
Similar to other models in the OSVP (e.g., IVI-RA model), IVI views the MDD model as a laboratory to test novel methods and to improve the science and practice of value assessment in the U.S.
The MDD model will allow users to evaluate the lifetime benefits and risks of various treatment sequences in U.S. adults (age 18-64 years) newly diagnosed with MDD by a healthcare provider, and to compare such evaluations from multiple perspectives (i.e., private and public payers, employers, people with MDD, and society). The model will feature both a health economic module and a multi-criteria decision analysis (MCDA) module.
How will the MDD model be different from traditional economic models?
Compared with similar efforts (e.g., a cost-effectiveness model used in HTA assessment), the IVI-MDD model is different in at least a few aspects:
- Its development is guided by continual engagement with a 20-member multi-stakeholder Advisory Group from the outset and throughout different phases of the development process;
- It seeks to incorporate direct patient input (e.g., patient preferences) in model design wherever possible;
- It enables users to test and compare the use of different input and modeling assumptions;
- It seeks to evaluate sequences of treatments (vs. a single treatment) that better reflect real-world treatment pathways;
- It is completely open-source – all model coding is available for replication and enhancement by developers, health economists, and others with interest in improving the science and practice of value assessment.
What is the model development process?
The flowchart below depicts our overall modeling process. In line with IVI’s commitment to an open and transparent model development process, we solicit continual input from a multi-stakeholder advisory group throughout the modeling process and public comment at every major stage shown below. To learn more, contact us at [email protected].
What is the target population for the MDD model in the draft protocol?
The target population is treatment-naive adults (aged 18-64 years) diagnosed with major depressive disorder. The current target population is based on input from the MDD Advisory Group and the initial model scoping review.
- The MDD Advisory Group highlighted the need to model the entire treatment pathway once the patient is formally diagnosed, given the potential for delays in initiation of treatment or “under” treatment.
- The initial version of the model will focus on the population aged 18 to 64 years, partly due to interest in impacts on productivity.
While IVI had initially considered focusing on individuals with “treatment-resistant” depression, there is no consensus definition of this state in the literature or in clinical practice. However, the flexibility of the model set-up will allow us to adapt it for other patient populations, which IVI intends to explore in future extensions. For example, the model will include the flexibility to evaluate outcomes for subgroups who have failed specific numbers (or types) of treatments.
What do you mean by “treatment-naive” individuals?
Treatment-naive is a clinical term referring to a person who has never undergone treatment for a given disease. While the patient population at the start of the simulations will be “treatment-naive,” the model set-up is designed to allow users to evaluate the long-term outcomes for patients that have failed a certain number/types of treatments.
Will I be able to use the model with my own data?
Yes. Users can easily modify model inputs through either the user interface or through the underlying source code. We welcome your feedback on the key model inputs that you would like to explore in the model and possible partnership opportunities to apply the model using data sources from you or your organization.
Will I be able to use the MDD model to model populations such as the elderly or those with co-occuring conditions?
Users can easily adapt the MDD model to generate useful insights for the subpopulations of interests that do not exactly align with our target patient populations (e.g., by using the inputs specific to the population of interest). However, it should be noted that some key considerations for such populations might not have been explicitly modeled in the initial version of the model. For example, the initial version of the model will not explicitly model the impacts of MDD treatments on other co-occurring conditions (e.g., diabetes, cardiovascular disease). IVI intends to explore modeling other subpopulations in future versions of the model, and welcome opportunities to collaborate in such efforts. Comments on target conditions and examples of related modeling efforts are welcome in this process.
How can the model be used?
The MDD model is designed to be a flexible, rigorous, and open-source model that can aid different stakeholders in a wide range of decision contexts. Section 6.8 of the model protocol describes some of the key outputs from the health economic module, and Appendix H describes some of the key research questions and applications the model can help address. We welcome your feedback on the key decision questions or contexts the MDD model can help inform, and look forward to partnering with your organizations in applying the model.
What kind of feedback would be most helpful from people living with MDD for the model protocol?
While the model protocol is intended as a technical document, we see multiple opportunities where people with MDD can provide feedback:
- Scenarios in which individuals switch treatments (section 22.214.171.124)
- Whether the key model inputs capture the impacts MDD and its treatments have on those living with MDD (section 6.8)
- Whether the proposed model outcomes comprehensively capture impacts important to those living with MDD
- Other issues that are important to those living with MDD that are not adequately considered in the model protocol
In section 6.8 of the model product, IVI has identified a list of key model outputs (or outcomes) from the simulations (see breakout box below). We hope that all people and organizations that comment on the protocol will let us know if we are missing an outcome or if there are other data sources they would recommend. We are especially interested to hear from people with lived experience if these outputs adequately capture the outcomes they have experienced.
The following outcomes will be tracked and counted to enable reporting and comparison across treatment sequences:
- Number of responses/remissions/relapses/recurrences
- Duration of response (number of cycles/months in response health state)
- Number of MDD-related hospitalizations
- Number of all-cause hospitalizations
- Number of suicide attempts and suicides avoided
- Life years (see below; may be more meaningful metric if suicide is included in the model)
- Quality-adjusted life years (QALYs)
- Costs (by specific category, such as MDD treatment, outpatient, inpatient, indirect, and total)
- Cost per clinical outcome (e.g., cost per response, cost per remission)
- Cost per QALY
Why does IVI include the QALY as a model output?
Including QALY as a metric in the model, along with many other key outcomes, will allow the flexibility to understand and evaluate the importance of looking at a wide range of outcomes and will allow comparison with prior economic evaluations that have used this metric. While the QALY is commonly used metric in existing economic evaluations, it has limitations and is considered by some to be discriminatory to certain patient subgroups (e.g., those with disabilities). Since any metric has strengths and limitations, IVI believes that it is important that decision makers not base decisions on any single metric (such as cost per QALY), but to consider a set of diverse clinical and economic outcomes for decision-making.
Why is IVI building an MCDA module in addition to a health economic module?
Multi-Criteria Decision Analysis (MCDA) is an alternative approach to economic modeling that allows value assessors to consider a wider range of criteria and priorities. As part of IVI’s effort to test and optimize use of novel methods, we hope this module will allow us to have a more nuanced approach to economic modeling, which is particularly relevant for such a diverse population.
How will IVI synthesize and prioritize the feedback received?
IVI will review all comments received and will post a copy of the comments on our website. In addition, we will consolidate comments based on topic area and themes. We will then discuss the comments with the MDD Advisory Group and the model developers. We will modify the protocol based on the following:
- New data sources or multiple recommendations to change an approach
- Recommendations and suggestions from consultations with the MDD Advisory Group
How does IVI intend to incorporate patient perspectives in building this model?
IVI sees the inclusion of perspectives of people living with major depressive disorder as a key objective of the model. As such, we have sought to ensure that both the overall approach and the outputs reflect people’s lived experiences in the real world. The MDD Advisory Group includes representatives from patient organizations, and we have consulted with the IVI Patient Advisory Council throughout the development process.
In addition, IVI is collaborating with the University of Maryland’s PAVE Center in a component project to understand what people living with MDD value most in treatments of MDD, and how they make trade-offs in selecting among different treatment options. We will seek to incorporate such input into different aspects of the model design to ensure that the model is patient-centric.
How can I stay informed of the development of this model?
By responding to IVI’s public comment period, we will continue to keep you updated on our progress. All comments will be posted on the IVI website, and we hope you will continue to provide feedback when we launch the model later this year.
How can we get in touch with the IVI team if we want to share proprietary data sources or apply the MDD model?
Please contact Dr. Richard Xie ([email protected]) if you would like to share data with IVI for the model. If you have additional research or recommendations, please share that information through [email protected].
How can I provide feedback?
Please submit comments either on letterhead via email at [email protected] or through our online survey form, which allows you to submit comments related to the key questions or to provide general feedback. No anonymous comments will be accepted, and all comments will be published on our website.