Wednesday, August 19, 2015

Structured Risk Benefit and Flibanserin

One of the most-visited articles on this blog is a 2014 article on structured risk-benefit, which makes some comparisons with how regulatory bodies are trying to cope with uncertainty and risk/benefit and how this might give insights to payer coverage decision-making (here).  The FDA's approval of Addyi (flibanserin) raises an opportunity to look at structured risk-benefit theory in the context of a drug approval that was a rapid media sensation.

"Addyi is a serotonin 1A receptor agonist and a serotonin 2A receptor antagonist, but the mechanism by which the drug improves sexual desire and related distress is not known."     [FDA press release, 8/18/2015]

More after the break.

The FDA approval of Addyi (flibanserin) for female sexual dysfunction provides an opportunity to review how structured risk-benefit evaluations might have helped explain this decision. Absent such an approach, the media seems to have perceived the new approval adversely.   Twice before, flibanserin had been unsuccessful in applications for FDA approval.  In the first effort, there was concern the "benefit" was not enough.  In the second attempt, the "benefit" (defined by clinical trial endpoints regarding sexual encounters) was balanced against risk (which here, as in most cases, is composed of completely different factors, such as hypotension and drug-drug adverse events) in a way that also denied approval.  Clearly, the "benefits" and the "risks" are essentially incommensurate and yet risk-benefit protocols require them to be weighed against one another.   This is seemingly an area where structured risk-benefit could have been helpful...or is it an example of the impossibility of much progress in the pathway to structured risk-benefit analysis?  Is the benefit 1.3 sexual encounters a month?  Is it the woman's sense of herself and an important part of her life?  Is it the overall context of a marriage succeeding or falling apart?  Should we think of this data in memory of a first date dinner we once had where the spark and excitement of the emotion/sexual harmony was breaking out fast and simultaneously for both parties?  How does that context fit with the famous thesis of evidence-based medicine that we should throw away the pages which are the introduction and conclusions of a published study and look only at the data tables when we evaluate its validity?  The FDA's phrasing to the advisory committee of experts was to: "Comment on the clinical significance of the observed placebo-corrected treatment effects of flibanserin on satisfying sexual events, sexual desire, and related distress."

The primary endpoint for Addyi was a statistically significant increase in satisfactory sexual encounters per month -  more difficult to evaluate by a placebo effect that was as larger, in absolute terms, than the drug effect (+1.5 events for placebo, with an additional +1 for drug).   On the risk side, the drug, which is taken daily, is not to be taken concurrently with alcohol, causes sedation, and has a significant range of drug-drug interactions.  On the other hand, the drop out rate in clinical studies was low (about 5% for placebo, and only about 10% for active drug) suggesting that the perceived side effects were manageable for 90% of patients.  Like most drugs, it is likely that some benefit quite a bit more from Addyi than others do.  Addyi's instructions explicitly recommend discontinuation is indicated if there's a lack of self assessed benefit at 12 weeks.  Finally, Addyi's approval has an additional guardrail in that it requires  a physician training program that precedes the physician's authority to prescribe.

On the morning of August 19, CNN was running a quite reasonable televised interview with a physician at Rowan University discussed the pro's and con's of Addyi.  

However, the media seem to have a surprisingly high rate of negative articles.   I provide some entry points to the recent press and also to some of the underlying FDA documents.   There has been no media attention to a pivotal event in the pathway towards approval, an FDA public advisory workshop on the general topic area (not the specific drug) in October, 2014.  Links to that meeting are also provided below.

According to an 800 page study by Harvard's Daniel Carpenter in 2010, the FDA is highly conscious (or behaves as if it were highly conscious) of its reputation in the media (see here for the book, here for the author.)   FDA did not have to approve Addyi; this is a decision of its staff experts although it is usually concordant the advisory panel.

NPR, February 16, 2015 (here).
"Female libido pill fires up debate about women and sex."

Slate, February 18, 2015 (here)
"Feminists claim the FDA is blocking 'Female Viagra' because of sexism.  They're wrong." 

WSJ, June 2, 2015 (here)
"FDA questions risks of drug to boost women's libido"

Bloomberg, June 5, 2015 (here)
" 'Female Viagra' may be more marketing than science."

Vox, June 19, 2015 (here)
"Five reasons to be skeptical of the new 'female Viagra' "

Forbes (June 28, 2015) (here)
"You can call it the 'female Viagra' but the libido pill, Addyi, won't come close to Viagra's impact."

WSJ, June 30, 2015 (here)
"Mediocre drugs raise costs for all"

FDA, August 18, 2015 (here)
Press release: "FDA approves first treatment for sexual desire disorder."

Washington Post, August 18, 2015 (here)
"From 1952-2015: The path to 'female Viagra' has been a rocky one."

The Guardian (UK), August 18, 2015 (here)
"FDA approves 'female Viagra' pill Flibanserin after two rejections."

Buzzfeed, August 18, 2015 (here)
"How Big Pharma used feminism to get the 'female Viagra' approved."

FierceBiotech, August 19, 2015 (here)
"The FDA blundered badly on the Addyi approval."

CNN, August 20, 2015 (here)
"Female libido pill maker Sprout bought by Valeant for $1 billion"
Also: NY Times (here)

IBJ, August 20, 2015 (here)
"Anthem to cover libido pill, other insurers weighing decision."

Financial Times, August 20, 2015  [Alphachat Podcast, here]
"Problems for 'female Viagra'. "

Business Insider, August 21, 2015 (here)
"How a tiny company brought the first 'female Viagra' to market against all odds."

Healthcaredive, August 26, 2015 (here)
"Insurers debate coverage for female libido pill."

JAMA, September 1, 2015 (here)
"Evaluation of flibanserin: science and advocacy at the FDA."
Opinion on JAMA article at MadInAmerica, here.

Bloomberg, September 17, 2015 (here)
"Female libido pill caused dissent in FDA ranks, memo shows"

FDA, September 2015 (here)
Decision Memo.

NEJM, December 9, 2015 (here)
Joffee HV et al. [FDA authors] FDA Approval of Flibanserin.  NEJM (epub 12.9.2015).

BioPharmaDive, December 11, 2015 (here)
Sprout CEO Whitehead heads for the exits amid lackluster Addyi launch.

JAMA Internal Medicine, February 29, 2016 (here)
Jaspers et al., Meta-analysis of data.
Also: "Debate Flares Anew," Medpage Today, here, in which some criticize Jaspers as a biased and subjective study.

Pharma DIVE, March 28, 2016 (here)
Sprout investors allege Valeant overpriced Addyi.

July 2015
Book:  Big Pharma, Women, and the Labour of Love, by Thea Cacchioni, here (Amazon).  An academic author, the has testified on flibanserin before the FDA. 

Homepage for the FDA's October 2014 two day workshop on "Patient Focused Drug Development: Female Sexual Dysfunction" (here.)  Videos of the workshop are also archived on this site, as well as PowerPoints and transcripts.  There is no question that this workshop was a key step in the FDA's progress toward approval of Addyi, even though it got drastically less media attention than the June 4, 2015 advisory committee meeting and the August, 2015 drug approval.  I have also loaded all the October, 2014 archival documents found on the FDA website into the cloud as a 5 mb zip file (here).  The workshop was actually part of the FDA's Congressionally mandated multi-year patient engagement initiative for better patient input into structured risk/benefit design (see here, here, here.)   However, it is difficult to see any trace of  the hoped-for "structured risk benefit analysis" that got out into the public media.

The homepage for the June 4, 2015 FDA Advisory Committee meeting, here.
   For the day's listing of all FDA meeting documents including panelists and agenda, here.
   For the FDA's systematic review of flibanserin (136 pp), here.
   Sprout Pharmaceutical's systematic presentation of flibanersin (291 pp), here.
   For 21 mb of FDA documents in one zip file in the cloud, here.
   As of August 31, the day's transcript isn't posted yet.


Two healthcare organizations asking FDA to not approve Addyi by formal letters, posted at (here).  After the June 2015 advisory board, an expert group wrote to the FDA as follows:

We believe that the 0-18-6 advisory committee vote on June 4 occurred because of a confusing scientific picture and the interference of “Even the Score,” a carefully orchestrated public relations campaign initiated by the makers of flibanserin and other companies with sexual pharmaceuticals in the pipeline. 
Approval would set a dangerous precedent and send the wrong message to the public that a drug for chronic use with a poor risk-benefit profile can be successfully promoted by unethical tactics. 
The discussion after the vote revealed the extraordinary ambivalence on the part of the B (approval with reservations) voters. For example, K. Curtis: “I voted B, but a somewhat conflicted and still uncomfortable B” V. Lewis: “I voted B, a difficult B” T. Gerhard: “I also voted B; a very difficult B, definitely; between a B and C” J. Perrone: ”I voted B, but I applaud the people who voted C” T. Stürmer: “I voted B. I was on the fence here, I have to admit.” E. Bell-Perkins: “I voted B; it was difficult.” W. Gellad: “I have serious, serious, serious safety concerns.” 
We will leave the topic of flibanserin’s safety to others, except for mentioning the truly absurd situation of approving a daily drug to boost the sex lives of women in their 30s and 40s that must not be taken with alcohol. As sexologists we can say with confidence that this advice is both preposterous and doomed. 
This is interesting because the panelists (as selectively quoted above) clearly were referring to two axes of decision-making, (A) what was decided, and (B) how certainly or strongly the decision was held, being very weak on the latter axis.  This topic of two-axis decision analysis is discussed in the September Harvard Business Review by Tormala and Rucker (here).  The level of rhetoric is high; obviously the FDA staff would not have thought their approval was either preposterous or doomed.