Dear Hope Now for ALS (HNFA) supporters,
We apologize for the delayed update. We have been trying to schedule the next meeting so we could report that in this update. After our July 6 meeting, we submitted a detailed proposal to the FDA and requested a timely response, but it appears the FDA is stalling. They suddenly want to invite other “stakeholders” to our meeting. We believe this is a tactic to bring people and organizations into the discussion that will defend the status quo and to convert the role of actual PALS and CALS into a minority voice. With much regret, we need to report that after three meetings we see little evidence that the FDA is willing to be flexible with respect to clinical study designs and data analysis approaches that could accelerate the development and delivery of safe and effective medicines for ALS. WE WILL NOT ACCEPT THIS!
Please write the FDA today and copy members of Congress and the media expressing your outrage that ALS is still not being treated with urgency. Demand that the FDA respond immediately to Hope Now for ALS (HNFA) in a substantive and meaningful way as they promised they would. State that you regard Hope Now for ALS as your representative for this process and that you expect the FDA to continue timely, direct discussions with HNFA on expedited and modernized development and approval pathways for ALS treatments – a 21st Century Approval Process!
Please email Dr. Woodcock at the FDA: email@example.com
and copy the people below and your representatives and media contacts
Summary of the July 6th conference call
Our third meeting with Janet Woodcock and her staff focused on finding common ground; specifically, balancing the need for faster, less costly, more conclusive trials and the FDA’s desire to keep ineffective treatments off the market. We were able to agree on FVC (change) as a primary endpoint, and ALSFRS (change) as a secondary endpoint. The ensuing discussion focused on topics fundamental to designing quick confirmatory studies that could lead to approvals for drugs/treatments that have demonstrated safety and preliminary evidence of efficacy in Phase 1 or 2 trials. Two principal topics that emerged from this discussion were (1) use of a patient-specific ALS progression predictive algorithm to construct a virtual and homogeneous placebo trial arm designed to reduce statistical noise in trials caused by the heterogeneity of ALS; and (2) understanding where and how the FDA can show some flexibility in the use of its statistical significance standard based on calculation of a p value as the standard for establishing “efficacy.” Achieving a p-value of < 0.05 in a trial where measuring a modest improvement in ALS progression is the primary endpoint necessarily increases trial size to substantiate positive outcome, adding time, cost and other complications that function as disincentives for companies to pursue approval of new medicines for ALS. However, by incorporating progression prediction models and patient-specific data, a 21st century approach to trial design, we would be more likely accurately measure the effect of the drug, improving the chances for recognizing of success when it occurs. We also advocated for the collection of secondary data in trials especially investigational biomarkers, and stratification of patients to better discern evidence of “efficacy” and to identify subsets of patients who benefit from a drug when efficacy is not seen in a majority of the patients. Unfortunately, as with our previous in-person meeting and our prior teleconference with FDA, we did not receive clear, actionable responses from FDA, and on some critical points of discussion, we encountered firm resistance to any form of new thinking in how drugs for ALS are developed and approved.
Detailed Notes of the July 6th Conference Call
On June 6, 2015, Hope Now for ALS held its third meeting with the FDA to discuss creation of an expedited development and approval pathway for new ALS treatments.
We opened the meeting with some sad news, informing the FDA that a member of our leadership team, Dr. David Huntley (husband of Linda Clark – also on our leadership team) lost his battle with ALS on July 4th. We informed the FDA of his passing at Linda’s request to remind everyone that making and delivering progress for ALS is a matter of extreme urgency. FDA responded with an expression of sorrow and a statement that they understand and share the urgency felt by the patient community.
Bob Hebron led the meeting for Hope Now for ALS (HNFA). The primary participants for FDA were Drs. Janet Woodcock and Ronald Farkas, although most of the FDA participants from the prior call also attended.
HNFA introduced the concept of using a predictive algorithm to select patients for a study, and then to serve as a second control arm alongside an actual control arm with the objective of reducing the statistical noise caused variability in the rates and characteristics of progression among PALS. The predictive algorithm approach was then discussed in some detail, including input from experts developing a predictive algorithm for ALS using existing patient data. It was explained that the algorithm accurately predicts decline in ALSFRS and can be adapted to predict other measurement endpoints, such as FVC.
HNFA then proposed that a clinical study designed to test for a slowing of progression could be conducted by using the predictive algorithm to select fast-progressing patients (termed “enrichment” by FDA) that could fill the study with “like” patients in terms of rate of progression. A positive effect on fast progressors might be seen more quickly (statistically), and in a smaller study population, than trying to discern the same effect in a population of mixed rate progressors. FDA indicated that enriching the study with fast progressors might be acceptable to them, and they expressed an openness to using the predictive algorithm for that purpose; however, they noted that if the study were populated only with fast progressors, the results might not be as generalizable to the larger ALS population as it would be if the study were run in a broader population of PALS (i.e., fast and slower progressing PALS).
HNFA then proposed that the study could perhaps be stratified (meaning looking at subgroups within the population) to learn more from the study, and to try to further reduce the statistical noise generated by the variability in the progression rates and types of motor functions lost as the disease progresses, seen at the individual patient level. FDA responded that attempting to stratify a study prospectively would complicate study design, lengthen the time needed for enrollment and cause the study to have to be larger (i.e., FDA would be looking for a much larger, longer study with each subgroup large enough to have adequate statistical power if the sponsor intended to draw conclusions regarding drug effect based on any subgroup analysis). FDA then said that retrospective analysis could be conducted on the study results to try to understand what happened, but the implication was that retrospective analysis would likely not support an approval.
HNFA then made the point that the patient community is looking for an agreement with FDA on smaller, smarter studies to keep the cost down and speed up the entire process. FDA’s response was entirely statistics-based and summarized below:
- A study can be small if there is a huge treatment effect
- A small study risks statistically missing a small treatment effect
- FDA reported that some Breakthrough drugs approved based on early data didn’t show a large treatment effect in later confirmatory studies.
- FDA advised that once an expected treatment effect is estimated (e.g., how much more slowly a drug might cause ALS to progress in the study population), cut it in half and design the study to be large enough to detect the smaller treatment effect to a level of statistical significance.
FDA did not directly respond to our statement regarding an agreement on smaller, smarter studies.
[Explanation – not from the FDA. What FDA actually said here is that they expect any entity trying to develop a new drug for ALS to do conduct the same large, time-consuming. randomized controlled studies today that the agency would have required 10, 20 or even 30 years ago. The FDA’s response is rooted in its long-standing its adherence to a purely statistical approach to designing studies and evaluating study results, and is based on the results of at least one, and preferably two, randomized controlled studies producing positive results at a measure of statistical significance, called a p-value, of 0.05 or less. Setting the p-value requirement at 0.05 or less means (in simplified terms) that the chance the results seen in the study happened because of chance (and not because of the drug) is 5 percent or less. FDA was explaining that an acceptable p-value can result from a large treatment effect in a small population of patients; however, to detect a small slowing of progression to a level of statistical significance acceptable to the FDA (p = 0.05 or less), a much larger, longer, more expensive study would need to be run. This convention, which has been relied on by FDA for decades as its standard for approving new drugs (since the 1960s), creates a reality in which getting a modestly effective drug approved by FDA takes much longer, costs much more and is much more uncertain for sponsors, than getting a highly effective drug approved. Given the current state of the medical science for ALS, all the drugs currently in the pipeline appear to have the potential for causing modest positive effects on disease progression and symptoms; thus, the FDA’s position on study size and statistical significance creates a financial and technical barrier to sponsors who might otherwise be interested in trying to navigate the agency’s development and approval process if a more creative and financially-feasible pathway was available. It also creates an obvious problem for the current and future generations of PALS and CALS who are faced with a drug development and approval process that makes approval of a new drug in a timeframe meaningful to them, all but impossible. HNFA is attempting to negotiate a more reasonable and productive development and approval path for ALS treatments with the FDA. So far, it appears that FDA has no interest in joining us in thinking outside the traditional statistical box.]
HNFA raised the concept of using the predictive algorithm to analyze the study results as a second control arm (run alongside an actual control arm with patients receiving a placebo) that might be a better way to control a study because of the algorithm’s potential to reduce statistical noise associated with variability among patients and the accuracy/precision noise known to be associated with measures like the ALSFRS and FVC. FDA replied that they would not object to looking at its use as a control arm, and that the usefulness of the algorithm could be evaluated by comparing the control outcomes predicted by the algorithm with the outcomes from the actual control arm. FDA described use of the algorithm for this purpose as exploratory, and appeared to be disinclined to rely on anything learned from its use to interpret the study results.
HNFA then asked: What is different about the input we are receiving from FDA today from what we would have heard 5, 10, 15 or even 20 years ago? FDA responded that the difference is they are talking to us about it – something they would not have done in the past. FDA did not offer any other explanation of differences between the advice they are offering us now versus what they would have advised in the past. HNFA then proposed that the study should include a cross-over provision (to the active drug) for patients randomized to the placebo control group, if after an appropriate period of following the patients in their respective groups (treatment and control), the drug appeared to be having a positive effect for some patients. FDA responded that they thought we were talking about a delayed start trial where some patients would start on a placebo, then be switched to the active drug. We responded that we were actually proposing a more traditional randomization to treatment or placebo at the start, with cross-over at the end of the monitoring period, and that if some of the patients who crossed over to the active drug then responded to treatment (for example by seeing their progression slowed), those responses should be considered favorable and could be compared to the predicted rate of progression using the algorithm. FDA did not provide a clear response to this concept but seemed to be discouraging it, and we raised the question regarding why FDA had never developed a way to consider patient responses observed after cross-over from the control to the treatment group as evidence supporting efficacy of the drug being tested? FDA responded that they were talking about these kinds of concepts at a high level but had not begun applying them at the study design and review level.
HNFA expressed its growing frustration that we do not seem to be getting clear and actionable responses from FDA. In response FDA asked that we submit a list of things the FDA had expressed “openness” about so they could confirm or clarify their position.
FDA reiterated its concerns about trying to use biomarkers as supporting evidence in an ALS study, stating that they had been burned by biomarkers too many times. (We did not pursue this discussion further since based on previous discussions, use of biomarkers at this time appears to be off the table with FDA. It seems they have no objection to collecting biomarker data – but they are not open to using them for purposes of regulatory decision-making regarding an ALS treatment.)
On endpoints, FDA stated that they think there is significant variability in ALSFRS and FVC measurements, and stated that if FVC is selected as the primary endpoint, it would not be evaluated in a vacuum. They would be looking for concordance in other measures (e.g., they would want to see ALSFRS and other patient outcome measures generally go in a positive direction if FVC outcomes were statistically positive).
During the meeting, HNFA expressed the following to FDA and FDA provided the following responses:
- We have made clear to FDA the ALS community’s perception of risks versus benefits, and that the community considers the risks of the disease to substantially outweigh the risks of investigational treatments that, based on early data, indicate they may provide some benefit.
- FDA responded that they understand that the ALS community supports approval of treatments with adverse side effects if the drugs also provide some benefit.
- HNFA clarified that the community extends their position on risk to FDA being open to approving drugs in an expedited manner based on limited data, and then relying on post-approval studies and registries to learn more, faster about the safety and efficacy of the treatment and how to optimize its use, even if granting the early approval might result in some treatments ultimately being found less safe or effective than originally thought, requiring re-labeling or even withdrawal from the market. Presently, PALS have almost nothing available to them and the pipeline is thin in part because of the high regulatory hurdles for ALS treatments; problems that can only be addressed through a stream-lined and expedited development and approval path. A specific recommendation that was made during the meeting was that FDA should be more flexible with its statistical significance standard (p = 0.05) for diseases like ALS with an extreme unmet medical need, which would allow for smaller, faster studies that, if positive at a lower level of statistical significance, could lead to a conditional approval (which is what accelerated approval already is), supported in the post-approval space by continued study.
FDA responded in a noticeably rigid manner that it must meet its statutory standard in the Food Drug and Cosmetic Act (FDCA), and that if the ALS community wanted the standard changed, they should pursue that with Congress.
[Explanation – not from FDA . The standard for approval in the FDCA is generally worded and allows broad discretion to FDA in deciding what the standard for approval should be for any drug intended to treat any disease. Congress has directed FDA to be more flexible when setting the standards for new medicines intended to treat serious and life-threatening diseases with unmet needs. ALS is among the most serious of all known diseases, and is almost universally fatal within 2 to 5 years) In our case, FDA has firmly indicated that they do not intend to be flexible.]
HNFA responded that FDA has broad discretion in interpreting and applying the approval standards in the FDCA, and that Congress had already instructed them to use that discretion for new medicines intended to treat diseases like ALS.
HNFA explained to FDA that our goal is to complete these discussions regarding a streamlined development and approval path for ALS drugs in a few additional weeks, then move to the next stage of proposing the pathway to sponsors for comment and potential use in starting registration clinical studies within the next few months, with the goal of completing the data collection portion of those studies by the end of this year, and that we expected FDA to partner with us in achieving this goal.
In follow up communications between HNFA and FDA since the July 6th meeting, HNFA has opted to submit a conceptual study design to FDA, and FDA agreed to provide feedback within approximately one week. The conceptual study design was provided to FDA and we expected a response this week. Instead, we received an e-mail from FDA stating that they would “…like to broaden the participants so more members of the ALS community can participate and provide input.” This decision by FDA, which was made unilaterally without consultation with HNFA, appears to mean that they do not intend to respond to our proposal in a timely manner. It also means that the pace of discussions with FDA regarding the modernization and acceleration of new drug development programs for ALS will be substantially slowed, and will lose focus as “stakeholders” with goals that differ significantly from the urgent needs of PALS and CALS are selected by FDA and brought into the process.