An Overview of Benefit-Risk Assessment in Drug Development
Rebecca Noel, DrPH, MSPHGlobal Benefit-Risk LeaderEli Lilly and Company
May 2017 (Proposed Presentation FDA DILI Conference XVII Session II)
Presenter
Presentation Notes
# 1. So good afternoon. Thanks for coming back. To the last session today. My presentation will be a bit of an odd duck in the context of this overall meeting in that I am not a hepatologist; I'm an epidemiologist by training, and for the last 10 years, have been specializing in benefit-risk assessment. My aim today will be to provide you with an overview of the thinking that we're using in industry and in regulatory agencies as we approach this topic of benefit-risk assessment.
• Employee and share holder of Eli Lilly & Company• The views and opinions represented in this
presentation are solely those of the presenter and are not intended to represent the views and opinions of Eli Lilly and Company
Presenter
Presentation Notes
# 2. By way of disclosure, I am an employee and shareholder of Eli Lilly & Company, and these are my personal views, not those of the company. So if you go away with nothing from my presentation other than these three things that I'm about to tell you by way of a high-level summary, I would be more than happy.
• Changing drug development, regulatory, policy and patient engagement landscape pointed to a need for increased consistency and transparency in benefit-risk evaluation
• Benefit-risk frameworks are a useful tool to guide and enhance decision making
• A simple, but not simplistic, framework composed of 5 steps
Presenter
Presentation Notes
# 3. And the first is that within industry and within regulatory agencies, we've been moving from an expert model of benefit-risk assessment to really using a structured approach to benefit-risk assessment. That structured approach is encapsulated by what we refer to as a framework-based approach. And what you'll see today is a simple, but not simplistic, framework.
The Critical Question for Regulators, Industry and Patients
Across the lifespan of a
drug, we ask one fundamental question…
Do the benefits of the drug exceed the risks for
the indication and its
expected use?
Presenter
Presentation Notes
# 4. So I heard in the previous session, the fundamental question that we're all asking, regardless of whether we're hepatologists or epidemiologists. As we're working to bring new therapies to market, the one critical question that we're always asking ourselves is, do the benefits outweigh the risks?
Benefit and Risk: Pillars of Regulatory Decision-Making
To be approved for marketing, a drug must be safe and effective for its intended use
• The meaning of “safe” is not explicitly defined in the statutes or regulations that govern approval
• Recognizing that all drugs have some ability to cause adverse effects, the safety of a drug is assessed by determining whether its benefits outweigh its risks
• This benefit-risk assessment is the basis of pre-market and post-market regulatory decisions
Presenter
Presentation Notes
# 5. But how do we really go about answering that question? How do we answer the question, do the benefits outweigh the risks? In the context of regulatory approval, we think that in order for a drug to be approved for marketing, it has to be safe and effective. Well, what do we mean by safe and effective? Well, there are three pillars to approvability. The first is efficacy. The second is that the drug or the therapeutic product has demonstrable quality. And the third is that it's safe.
A benefit-risk framework is a systematic, consistent, and transparent approach (e.g., process and tools) that guides the assessment
Structured benefit-risk assessment uses the concept of quality decision-making via “frameworks”
Historic reliance on expert judgment, but moving toward use of expert judgment +structured benefit-risk assessment
Having said the safety of a drug is assessed by whether its benefitsoutweigh its risks, just how should we make that determination?
Presenter
Presentation Notes
# 6. But when you look in the regulations, you might be surprised to find that nowhere is safe defined. We define safe is that the benefits outweigh the risks. And this benefit-risk assessment is the fundamental piece of all regulatory and drug development decision making. You might've thought to yourself that this might seem to be a bit circular, and I'm not sure how I draw myself out of this circuitous loop of decision making. Well, we said in the previous slide, the way that we determine approvability is that the benefits must exceed the risks. Historically, we've assessed benefit-risk assessment through expert judgment. But we're moving towards a new paradigm in which expert judgment is supplemented with structured approaches to decision making.
B-RA Frameworks: Support for Decision-Making and Communication
EMAFDA
PhRMA BRAT
BRAT = B-R action team; EMA = European Medicines Agency; FDA = Food and Drug Administration; PhRMA = Pharmaceutical Research and Manufacturers of America
PRoACT-URL
Presenter
Presentation Notes
# 7. So what does that structured approach looks like, look like? It looks like higher quality decision making through the use of these decision aid tools that we typically refer to as frameworks. Now, I don't expect you to be steeped in this as are some of us in the epidemiology and benefit-risk community. But there are a variety of frameworks. So you can talk to us and we might mention the FDA framework, we might mention the EMA framework, we might mention the industry framework. And those are, these are some of the frameworks that you commonly hear referred to.
Five Decision Factors –Analysis of condition –Current treatment options –Benefit –Risk –Risk management
Two levels of consideration –Evidence and uncertainties –Conclusions and reasons
Summary –Benefit-risk summary assessment
Presenter
Presentation Notes
# 8. Let's do a bit of a double click, and look at the FDA framework. The FDA framework is composed of five decision factors. They're called the analysis of condition, current treatment options, benefits, risks, and risk management. Beyond the decision factors, there are two levels of consideration, evidence and uncertainties, and then conclusions and reasons. And really, this is, in terms of evidence and uncertainties, it's, what's your data. And then conclusions and reasons is what do you draw from that data. What does that data mean? And then this is underpinned by an overall summary assessment.
International Council on Harmonization (ICH): Presentation of Benefit-Risk in Submissions and for
Post-Marketing Surveillance
EMAFDA
PhRMA Brat
Structure & Presentation of Benefit-Risk Information
Presenter
Presentation Notes
# 9. I mentioned that if you approached a benefit-risk person in industry, they might rattle off several frameworks, and it might just sound like a giant alphabet soup. So how do we move from a alphabet soup to something that we can all align around, wrap our heads around, and begin to use in our respective work places? We do this, in industry and regulators, through an organization called ICH, the International Council of Harmonization. And this is really where industry and regulators come together to work out what we believe should be our common approaches to common problems. I'm going to talk about now is what actually came out of ICH. For those of you in industry, this is the framework that you will see your benefit-risk argument put into. On the other side, if you're in a regulatory agency, this is the way in which you can expect to see that benefit-risk argument come into you for review.
• Espouses a framework, not specific methods• Benefit-risk assessments is based on a weighing of
key benefits and key risks• Weighing implies judgment and allows for quantitative
approaches, where useful• “Key benefits” and “key risks” are those that
contribute most importantly to the benefit-risk profile• A descriptive (qualitative) approach is usually
sufficient• Critical to communicate the implicit thinking behind
what is assessed as being important and why
Presenter
Presentation Notes
#10. Before we decided how we would approach the concept of a common framework, we laid out some common guiding principles, and you see those reflected here on this slide. What we agreed to was that we weren't going to focus on specific methodology. So for those of you who know benefit-risk, you will know that even beyond the alphabet soup of frameworks, there's an even greater number of quantitative methods that can be used to produce a benefit-risk assessment. That's not the direction that we wanted to move in, because as we've heard here earlier today, one size does not fit all. So we wanted to step away from specific methodologies and really think about the, about the process. So this is a framework, not a specific methodology. Embedded in our thinking is the idea that a benefit-risk assessment is based on a weighing of the key benefits, the good stuff, and the risks, the bad stuff. And weighing implies the use of judgment, and allows for quantitative approaches where and when it's appropriate. This idea of key benefits and key risks is something that people can get tripped up on, but essentially what it is at the end of the day is, what are the things that are most impactful to the overall benefit-risk assessment?
• Why begin with a qualitative assessment?• Concern about reducing a complex decision into a
single, summary statistic• Some quantitative methods obscure expert judgment• Regulatory review is characterized as a qualitative
exercise grounded in the quantification of various data
• Provides the flexibility to accommodate quantitative analyses if warranted
Presenter
Presentation Notes
#11. Last but not least, you'll see that this framework embodies a qualitative approach to decision making, because we believe that the vast majority of benefit-risk decisions don't require quantification. Once you work through the process, it really is fairly self-evident whether the benefit-risk balance is positive. So why focus on a qualitative assessment? Well, as we came together as an expert working group, we shared the belief that you can't take a benefit-risk assessment decision and reduce it to a single summary statistic. There's also the recognition that quantification, or the use of some very specific methodologies, actually obscures expert judgment, and that is not what we want to do. We're looking to supplement, enhance, and make our decision making more transparent, not to obscure it. There's recognition that regulatory review is a qualitative exercise grounded in quantitative data where and when it's appropriate, and to support the ability of quantitative methods to grow and evolve.
#13. It took us about two years from 2014 to 2016. This overall graphic gives you a snapshot of what our thinking was. You'll see that any benefit-risk decision always occurs within a context, and we refer to that context as the therapeutic context.
Steps 1 & 2: Analysis of the Condition and Treatments Options
• Sets the stage by creating the context for the weighing of benefits and risks• How serious is the condition and why?• Who, how many and to what degree affected?• How well is the patient population’s medical need
being met by available therapies?• Focus on aspects with the greatest relevance and
impact (e.g., incidence, morbidity, mortality, QoL)• Public health and societal implications should also
be addressed where relevant
Presenter
Presentation Notes
#14. The therapeutic context is composed of an understanding of the disease or condition, plus a very thorough characterization of what the current treatment landscape looks like. Within this therapeutic context, you then assess the benefits and the risks, and you reach your overall summary conclusion. In the first, let's take a look at the analysis of condition and treatment options. So we refer to this as Step 1 and Step 2. As I mentioned, this really sets the stage by creating the context for the weighing of benefits and risks. So it helps us understand, how serious is the condition and why? It helps us understand who's affected, how many of them, to what degree? And it also helps us characterize how well that patient population's needs for therapy are being met. It also helps us focus on aspects with the greatest relevance and impact to the decision, and also incorporates an understanding of public health and societal needs, when and where it's relevant.
• A presentation of a factual summary of the data, i.e., without interpretation
• Typically primary and other clinically important endpoints (secondary and exploratory endpoints)‒May use combination of study endpoints‒Show ability of surrogate to predict clinical benefits‒May include convenience (e.g., dosing regimen or route of
administration) that may lead to improved patient compliance or benefits that affect those other than the patient (e.g., population benefits of a vaccine due to herd immunity)
• Discussion of strengths, limitations and uncertainties of benefit data
Presenter
Presentation Notes
#15. For the next step, Step 2, we need to assess the key benefits. So this is where we present a summary of the data without any interpretation. Typically, when we're looking at key benefits, we're looking at primary and other clinically important endpoints, secondary and exploratory endpoints. And in the context of presenting your factual data, there should also be a very straightforward discussion of strengths, limitations, and uncertainties of your benefit data. This sounds really straightforward. I need to talk about my therapeutic context, the disease, analysis of condition, what the treatment landscape looks like, and then I'm going to talk about my key benefits. I know that it might be my primary endpoint or my secondary endpoint, but beyond that is there really any guidance for what a key benefit looks like?
Identifying and Describing Key Benefits• Clinical importance of the benefit
• Magnitude of the absolute difference in effect vs comparator; in some cases, also expressing the difference in relative terms may be informative
• Time course of the key benefit
• Variability, particularly in relevant subpopulations
• Study design considerations
• Completeness
• Number of clinical studies and consistency of results across studies
• Dose-response
• Generalizability of the clinical study result to clinical practice
• Uncertainties around surrogate endpoints predicting benefit
Presenter
Presentation Notes
#16. Well, yes, there is some guidance for what a key benefit looks like. When we're asking you to think about what's most impactful to the decision making from the benefits perspective, we're asking you to consider things such as the clinical importance of the benefit. The magnitude of the absolute difference and effect versus the comparator, time course of the key benefit, Variability, especially across sub-populations, study design considerations, completeness of your data, dose response, number of studies and the consistency of the findings across those studies, generalizability, and then uncertainties around surrogate endpoints that might be used to predict benefit.
• Similar to key benefits, key risks are a presentation of a factual summary of the data, i.e., without interpretation
• Risks include adverse events and other unfavourable effects associated with the product
• Discussion of strengths, limitations and uncertainties of risk data
Presenter
Presentation Notes
#17. So let's flip over then to key risks. We've talked about the good stuff, now let's thing a little bit about the bad stuff, the risks. This is very similar to the presentation of key benefits, in that key risks are also presented in a factual, straightforward way without interpretation. Risks typically include adverse events or other unfavorable effects associated with the product, and to discuss strengths, limitations, and uncertainties.
Identifying and Describing Key Risks• Seriousness and/or severity, frequency, reversibility, and tolerability
• Frequencies as the absolute difference relative to the comparator in the context of the background frequency in the patient population
• Ability to monitor, minimize, or manage the risk
• Variability of the key risk across relevant subpopulations
• Time course of the risk
• Study design considerations
• Adequacy of assessment of risk
• Investigation(s) to address safety issues
• Completeness of data collection and duration of follow-up
Presenter
Presentation Notes
#18. So how do you think about identifying a key risk? What really contributes to raising a risk to the level of a key risk? Some of the things that you can use to guide the identification of a key risk include seriousness or severity, frequency, reversibility, tolerability, ability to monitor, minimize, or manage the risk, time course of the risk, adequacy of your risk assessment, any investigations that might've been done to characterize or discharge the risk, and then the completeness of your data collection, and the duration of followup.
• A benefit-risk conclusion, provided via a succinct explanation of the reasoning and judgment used in assessing and weighing the key benefits and key risks
• Explanation of how any uncertainties affected the interpretation of the evidence and their impact on the benefit-risk assessment
Presenter
Presentation Notes
#19. When it all comes together in this integration step, you provide an integrated benefit-risk assessment. This really is a conclusion provided via succinct explanation of the reasoning and judgments that went into weighing of these benefits or risks, as well as a thorough discussion of how uncertainties may have impacted the interpretation of the evidence.
Aspects to Consider in Reaching a Benefit-Risk Conclusion
• The impact of the therapeutic context on the BR assessment• How the severity of disease and expected benefit influence
the acceptability of the risks of the therapy• How the new medicinal product addresses a medical need
• Key aspects of risk management that are important in reaching a favorable benefit-risk assessment:• Whether nonresponders can be readily identified, allowing
them to discontinue treatment• Other risk management activities, such as registries or
restricted distribution systems • Remember: Methods that quantitatively express the underlying
judgments and uncertainties in the assessment can be used
Presenter
Presentation Notes
#20. So some additional aspects to consider in reaching a benefit-risk conclusion is really to think about how the therapeutic context impacts the overall benefit-risk assessment. We mean how the severity of the disease and the expected benefit may influence the acceptability of the treatment, and how this new product or the therapy is really addressing a medical need that exists among our patients. Another aspect to bring forward is this issue of risk management. There are certain situations in which a risk of itself might seem insurmountable, but in the context of a risk management program, might become quite acceptable.
Hepatic Toxicity through the Lens of Benefit-Risk: Real World Example
• Ribociclib was approved for treatment of postmenopausal women with hormone receptor (HR)-positive, human epidermal growth factor receptor 2 (HER2)-negative advanced or metastatic breast cancer• Grade 3 or 4 increases in ALT (10% versus 1%) and
AST (7% versus 2%) were reported in ribociclib and placebo arms, respectively
• Concurrent elevations in ALT or AST greater than three times the ULN and total bilirubin greater than two times the ULN, with normal alkaline phosphatase, in the absence of cholestasis occurred in 4 (1%) patients
Presenter
Presentation Notes
#21. Let's look at using the framework, at a high level, through the lens of a recent approval. Recently ribociclib was approved for the treatment of advanced and metastatic breast cancer. However, when you look at the data associated with ribociclib, you see that there was quite a bit of hepatotoxicity. There were Grade 3 or 4 increases in ALT and AST, along with concurrent elevations that were more than 3xULN. There were also actually cases of Hy's Law witin the approval package.
hepatobiliary toxicity• Four patients (1.2%) in the ribociclib plus letrozole• Group met the biochemical definition of Hy’s law;
• Despite the availability of hormone directed therapies for treatment of first-line Hr positive advanced breast cancer, patients ultimately develop resistance and progression of disease and go on to receive multiple additional therapies. In light of the high burden of disease, there remains a clear medical need to develop new therapies for the treatment of advanced breast cancer to extend life, delay disease progression and/or lessen breast cancer related symptoms. The risk of QT interval prolongation, hepatobiliary toxicity and neutropenia and dose modifications to address safety will be communicated in labeling in the Warnings and Precautions section of the label.
Presenter
Presentation Notes
#22. So we have this very significant safety issue, but let's step back and look. The key benefit in this case was progression-free survival. Key risks included QT interval prolongation, neutropenia, and the hepatic issue. But we put that all into the therapeutic context and the unmet need. This a very serious disease, and the ability to offer increased survival in the context of appropriate labeling and a risk management plan, led the FDA to conclude that ribociclib was indeed approvable
#23. This is a reference, for those of you who may like to look at the ribociclib information, there's a very nice description in this data package of the risk management plan that outlines just the thinking that we walked through.
• FDA Structured Approach to Benefit-Risk Assessment in Drug Regulatory Decision-Making: http://www.fda.gov/downloads/ForIndustry/UserFees/PrescriptionDrugUserFee/UCM329758.pdf
• Coplan PM, Noel RA, Levitan BS, Ferguson J, Mussen F. Development of a framework for enhancing the transparency, reproducibility and communication of the benefit-risk balance of medicines. Clin Pharmacol Ther. 2011 Feb;89(2):312-5. doi: 10.1038/clpt.2010.291. Epub 2010 Dec 15.
• Revison of M4E Guideline on Enhancing the Format and Structure of Benefit-Risk Information: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/CTD/M4E_R2_Efficacy/M4E_R2__Step_4.pdf
Presenter
Presentation Notes
#24. There're some additional selective readings, for those of you who may be interested in learning more about benefit-risk assessment.
#25. And I will conclude and wrap up with acknowledgment that even though I'm not a hepatologist, even though I'm just a lowly epidemiologist it doesn't matter whether you're a hepatologist, it doesn't matter if your expertise is in benefit-risk assessment. We're all here for the same purpose, and that is to help one another reach better decisions using all of our available information. Thank you.