Presented by Ari Gnanasakthy, RTI Health Solutions
So the purpose of the presentation today is really three topics. We’ll define labels. You know, we talk about labeling and claims and label, and these are all slightly different things. So let’s define labeling. And I’ll sort of touch on too, what’s labeling and what’s advertising as well, as much as I know, it’s a very complicated topic. Then we will review the labels for the new drug applications, not all drugs but simply the new applications for the last five years, 2011-15, and then we will compare them with the labels approved for 2006-10, so the last five years versus the five years before that. So we have data for ten years, in fact, so we’re looking at the last five versus the previous five.
Good, so what is labeling? Labeling is also known as various things in the industry. It could be prescription information, package insert, it is also called professional labeling, direction circular or package circular, it all depends on who you talk to within the regulatory or within the medical field. So labeling is called various names.
And the general requirements for prescription drug labeling—because there are device labeling and nutritional labeling and all sorts of different labeling, so this is prescription drug labeling—it must be a summary of efficacy and safety, use of drug. Okay, so safe and effective use of the drug. It must be informative and accurate—that kind of makes sense. It must not be promotional, false, or misleading, okay, so you can only say what was proven, you can’t take it any further. You can’t say, for example, if the drug is only for lowering cholesterol, you can’t say it lowers cholesterol and therefore you will live longer, or something along those lines. Same thing with cancer, if it shows tumor reduction, that’s all you can say. You can’t say it reduces your tumor and therefore you can live longer, okay. So it must not be promotional, false, or misleading. No implied claims or suggestions for use if evidence of safety or effectiveness. Based whenever possible on data derived from human experience. Okay, so that’s the drug labeling.
Labeling is any written, printed, or graphical material containing drug information and disseminated by or on behalf of the manufacturer of the drug, so that’ll be the sponsoring company that did the clinical trial. And for use by healthcare professionals. We sometimes mistake labeling, thinking this is for consumers. That is not the case. Labeling is for healthcare professionals. So the primary purpose of drug labeling is not to inform patients about their medications, it’s really for medical professionals.
So here is an example of a package insert, and you would have seen this whenever you open a prescription, you will see this information that you cannot read. I mean you don’t even read what goes on the bottle, let alone what’s on this package. But this piece of sheet—I was going to say something else—sheet—I have to be careful with my accent, and you know, speed and so on. Anyway. This sheet can go to pages, okay, it looks short now with lots of information—efficacy, safety, clinical trial information, contraindications, black box warnings, and all sorts of things. It can go into pages.
But advertisements, on the other hand, is any broadcast through audio, video, or visual media such as radio, television, and so on and so forth, okay. So advertisement is much broader and labeling is kind of everything else. Now guys, it’s much more complicated than this, and this is all I know. Labeling is very different from advertisements, okay. Labeling is very controlled. Advertisement takes a little bit further, so you can make some claims, but those claims must be agreed by the FDA. So here is an advertisement: “Help for fibromyalgia pain can start with Lyrica.” So this is, again, this is an advertisement, and all the words that you see on the top: “Help for fibromyalgia pain can start with Lyrica.” That’s a claim that was agreed between Pfizer and the FDA. And this advertisement, as you might guess, is based on patient-reported outcomes for pain and functioning.
So requirements for advertisement and promotional labeling. Claims must be consistent with the prescription information, so what you find in the package insert must be consistent with the advertisement. And it should not be false or misleading statements or omissions. Materials must be fairly balanced, benefits and risks. So you can’t just simply advertise what’s only the benefits, but you must also say what the risks are. And that’s why the printed advertisement that you see in People magazine or Reader’s Digest or whatever, it can go on to like two, three, four pages sometimes. And it must be based on substantiated data. So here is an advertisement, and this is nothing new. We have seen these. I have often fallen victim for those that claim six-packs in three days, kind of thing. Obviously it didn’t work, as you can see. But this is not based on a clinical trial or anything like that. So this is okay, we can let it pass. But here is one of the most misleading advertisements ever, frankly, with Brooke Shields for Latisse a few years ago. It was for improvement of eyelashes, but one of the side effects was that if the medication touches somewhere else then you start growing hair there too. And that wasn’t mentioned in the promotional material, so it’s supposed to be one of the most misleading commercials ever. So we are not going to rush out and buy this thing now at a local CVS or something like that, okay. Good.
And here is a recent commercial, again based on patient-reported outcomes for overactive bladder. Urgency, frequency, and leakage, a very controlled experiment, a good commercial, you will have seen them on TV as well as printed material. And this is the first of I think like three pages of advertisements, because they must balance the efficacy and the safety information.
And here is another one by Bristol-Myers Squibb for ORENCIA, for rheumatoid arthritis. And this is all based on, again, patient-reported outcomes, runs into pages, and as you can see this is based on the health assessment questionnaire that has items like can you open this, fasten this—shoelaces and so on and so forth.
So labeling is very different from advertisements. Advertisements are also very well controlled. Okay, I think we are okay with that. Good.
So now we switch to the real topic of the day, drug approvals between 2006 and 2010. So FDA approved 2016 new drugs between 2006-10—not ’11-’15, ’06-'10. And 24%, roughly a quarter, of the new drugs had at least one PRO-based labeling. Now, there was a publication before that, that was published long before the PRO guidance, that said there were about 33% of drugs, about one in three drugs, had at least one PRO labeling. So there was a reduction of labeling and this caused some waves, and we published this in 2012, a review of patient-reported outcomes labelings in the United States. So what I am about to show you is the repeat of this work but for ’11-1’5.
So drug approvals ’11-’15. So the total number of new drugs approved between ’11 and ’15 is 182. New drug approvals. The number of new drugs with at least one PRO labeling was about 30. And that’s about 16.5%. So the previous five years, it was 24%. And now it’s 16.5%. So 16.5 versus 24.1, obviously in terms of percentages, there has been a reduction between ’06-’10 versus ’11-’15. Now, so this is kind of disappointing for those in the PRO field, and especially within the regulatory field, because everyone is talking about patient-focused drug development and the PDUFA V encourages involvement of patients everywhere. So this was a bit of a surprise. So we wanted to dig a little further into the data. And one of the things that you would notice is that the 2011-15 was based on 116 approvals. And 2006-10 was 182. Okay, so we wanted to look into this a little further.
What we found was that there was a noticeable increase in the new drug applications for three disease categories. Cancer drugs went up by about 180%, infectious diseases went up by about 100% or 90%, and endocrine and metabolic diseases approvals went up by about 115%, while all others went up by about 6%. So on average, there was an increase of 136% between the first five years and the second five years. And most of it was driven by either cancer drugs, infectious disease drugs, or drugs related to endocrine and metabolic diseases. Now, what do we know about these three groups of drugs when it comes to patient-reported outcomes? Not a lot happens in terms of labeling, because cancer drugs are not dependent primarily on patient-reported outcomes data. Patient-reported outcomes are there, but it’s not used for regulatory decision making. Same thing with infectious disease. Most infectious disease applications are based on practitioner activities—CD4 counts, bacterial counts, and so on—not based on patient-reported outcomes. And endocrine and metabolic diseases like diabetes and so on, yes there are PROs but they’re not used for regulatory decision making. So it’s not surprising, in some ways, that the number of PRO labels have gone down proportionally.
So we went a little further. When I looked at the labelings for the three major diseases that we just talked about for ten years, ’06-’15, there were only five labels—about 3% or 3.5%. When we looked at all other diseases, there were 53 labels, about 34%. So this is based on about 300 applications—new drug applications—and about 60 labels. So the hypothesis, or the assumption, or the belief that these three drugs don’t yield much PRO labeling is kind of reasonable.
So here is a commercial. Here is an advertisement for an HIV treatment, so this is infectious—well you can say, infectious disease. “Our medication will not make you younger, sexier, or smarter.” Okay, so right off the bat they are saying, look, PROs are not the thing here, we don’t really care what you do with this thing. “Taken in combination with other HIV medication, we can lower your viral load.” So this is the promotion based entirely on viral load, okay. They are not saying, look, if you want to be the best in the world of this, that, and the other, that’s up to you. But we will get the viral load down, okay, CD4 counts down. Very proud.
But here’s another commercial for Kybella a few months ago. This was for double chin. It’s a very very good case study, this needs some writing up. The patient-reported outcomes was the core primary endpoint, and it was based on treatment satisfaction. Now, we always talk about proximal concepts and distal concepts, and we always think that the distal concept, you know, patient satisfaction is a distal concept and therefore there will be no labeling. But in this indication, what we usually consider as a distal concept was a proximal concept. Okay, the questions were things like, I look prettier, I look less older than my age, and things of that kind, where the content validity can be very questionable. But that’s what makes this product a success. Are you satisfied with the fact that your double chin is no longer there. So it was a core primary endpoint, a very good case study.
So there were two categories of diseases we can postulate, if you like. One that depends on patient-reported outcomes, or traditionally depends on PROs, to demonstrate treatment benefit. Traditionally. Going forward, that may not be the case, because lately there have been quite a few publications from the FDA insisting or sort of asking us to think about more PROs in cancer studies, for example, because patients with newer cancer therapies are living longer, and the differentiation between one therapy and the other is very little, and they want to see what the toxicity, what the side effects from the patient’s point of view might be. So going forward, some of these assumptions may not hold. But traditionally, there are drugs that are dependent on PROs for regulatory decision making. And there’s a whole bunch of others like infectious disease and endocrine diseases that are not dependent on PRO.
So what we did was to go through ICD-10 classifications and said, okay what are the sort of diseases that don’t traditionally use PROs for regulatory decision making. Okay, so there’s cancer, there’s infectious and parasitic diseases, endocrine nutritional and metabolic diseases. Diseases of the circulatory system, where we talk about chronic heart failure, MI, PHD. Then there are disease areas where there haven’t been many broad approvals, like pregnancy/childbirth, symptoms and signs, and abnormal and clinical and laboratory findings, injury/poisoning, suicidality, external causes of morbidity like transportation accidents and military operations. So these are the sorts of disease areas we thought that don’t depend on patient-reported outcomes for decision making. So we grouped them.
And there’s a whole bunch of others that depend on patient-reported outcomes for decision making. Blood and blood-forming organs. What happens when we have problems in blood-forming organs. We have fatigue. Fatigue, tiredness, that’s all these patient-reported outcomes. Mental behavior and neurodevelopment disorders, like schizophrenia or anxiety disorders. Diseases of the nervous system. Eye, ear, respiratory system, digestive system. Skin and subcutaneous systems. Musculoskeletal, like osteoarthritis and lupus and gout and rheumatoid arthritis. Genitourinary systems—erectile dysfunction, overactive bladder, dyspareunia, and so on and so forth. These are all pretty heavy PRO-dependent diseases.
So if you guys, if your company is working in these disease areas, you know, job security guys, you’re okay. But if you’re working on infectious diseases, maybe network later today. You might want to think about it.
So there was a noticeable decrease in the percentage of NDAs relating to PRO-dependent diseases. Okay, so when we look at the PRO-dependent diseases, in 2006-10—that’s the blue bar—there were 42% of drugs. But now, later years, there were 27% or thereabout. In the non-PRO-dependent, it was about 60% that went up to about 70%. So this is PRO-dependent and non-PRO-dependent. So it’s the same hypothesis as before, instead of looking at just cancer, infectious diseases, and the metabolic diseases, we sort of grouped the diseases by going through the ICD-10 code. And the assumption still holds.
So PRO labeling was comparable if we look at it between the two disease areas, if we take account of the PRO-dependent and non-PRO-dependent groupings. So if we take the first five and the second five years, the first five years there were 28 labels between 2006 to 2010. Between 2011-15, there were 30 labels. So the number of labels between these two groups of five years were about the same, okay, 28 versus 30. But the denominators were different, that’s what was driving the percentage difference, okay. But if we take the 23 of the 2006-10 for the PRO-dependent drugs, and the 23 for the 2011-15 on the basis of PRO-dependent drug, the percentages are about the same, about 47% versus 46%. So whereas before, it was 24% versus 16% difference between the two groups of years, but if we take into account what was actually approved, the type of drugs that were approved, the percent of labels hasn’t changed much. Now should they have improved, should it be more? Maybe so. That’s a different argument, I guess, or discussion. But it’s not. The reduction wasn’t there. It’s about the same.
So if we look at the PRO-dependent diseases between the entire ten years, and to see where things are happening. So if you have a drug store GI system—and I also have a number of drugs in the “n” parentheses as well. There was only one drug for things to do with ears, so we can leave that out. But nervous system, digestive system, musculoskeletal, blood-forming organs, these are PRO-heavy, if you like. Especially the musculoskeletal system, 46% of the drugs had a PRO label. Now, should it be 80%? That’s a reasonable argument. But we need to look into that a bit more further. Nervous system drugs mainly consisted of anti-epilepsy drugs, which rely on seizure rates reported by patients, okay so that’s pretty heavy. Skin and subcutaneous, the derm division within the FDA only recently recognized the value of patient-reported outcomes, okay, only since the Cosentyx was approved for psoriasis, they started taking PROs seriously. So we might see, when we look these same sort of figures in the next five years, it is not 40% anymore, it might be 70 or 80%. Things might change. So when we look at the non-PRO-dependent drugs, again looking at the entire ten years, you can see where things are.
These are all field drugs by the way, symptoms not classified elsewhere, this is usually symptoms due to chemotherapy. It’s not due to cancer but due to chemotherapy or some other sort of secondary symptoms. Injury and poisoning, there were only two submissions. But if you look at something like the endocrine, nutritional, and metabolic diseases, there were 41 submissions, only six—41 submissions that led to about four labels, about 10%. Infection and parasitic diseases, I think there was one drug, something to do with ear infection. Cancer drugs, there was hardly anything in the new drug approval. I mean, we published a paper a few months ago in the Journal of Clinical Oncology where we cited two labels for cancer drugs, but that’s not based on new drugs, that was a supplement of NDAs.
So in summary, overall labeling declined from 24.1% to about 16.5% from 2006-10 to ’11-15. The number of new drugs approved that traditionally uses PRO-related endpoints was about the same between the two groups of five years. The number. So about 49 and 50. And when we look at percentages, they were about the same too, 47% and 46%. Whether this is good news or bad news I’m not so sure. It’s also very difficult to write this as a targeted message in a manuscript, because there are different messages. People can take it whatever they want to. Oh it’s bad news because the percent of labels have gone down, that’s the attitude. But if you’ve missed the second bullet, then that would be unfortunate.
But there are some observations as well, and we don’t have a great deal of time to go through them in detail, but I’ll give you the bullets behind them. Almost all labelings were for proximal concepts, so proximal, by that I mean symptoms or function. Now I used to think that there was symptoms and then, sort of, problems with symptoms lead to problems with functioning, as two separate things. One thing I realized over the last few years was that there are many diseases where functioning is the first-order impact, not the symptom. So if we take for example, irritable bowel syndrome or impact of chemotherapy, going to the bathroom, diarrhea for example. I used to think that was a function, because it’s bowel movement, you know, it was a function. But actually it’s a symptom of that disease. So it’s a symptom. Now we can argue about that too. But what I’m trying to say is there is no such thing as sort of symptoms, function, and health-related quality of life and quality of life. There’s no—I don’t think there’s a continuum there, I think symptoms and functions are the same. But nevertheless, most of the labels were for proximal concepts.
There were a couple of exceptions. One was for Arcapta Neohaler from Novartis, that the label was for health-related
quality of life based on SGRQ, the St. George’s Respiratory Questionnaire. It’s what’s called a useless label. Anyone want to guess why that would be a useless label? Because nobody knows in the US who St. George’s is or what he got to do with anything to do with respiratory infections. And a four-point reduction in SGRQ, the company couldn’t promote it. There were quite a few labels that came out—I think there were two drugs that came out from the respiratory division that’s kind of what I call useless labels.
So some labeling had limited value to the sponsors. Most PROMs, or PRO instruments, were simple diaries or legacy instruments. It doesn’t mean just because they were simple diaries they wasn't any documentation or a great deal of content validation that went behind it. What I’m trying to say here is that they were not multi-dimensional measures with different domains and total scores and etc. Most were very simple diaries or legacy measures like the St. George’s Respiratory Questionnaire, which leads to a useless label, or a cystic fibrosis questionnaire that led to a very good label for Kalydeco, for cystic fibrosis.
New measures are not always publicly available. Now this is kind of interesting, especially for us, who are in the PRO domain, because we go to PRO Consortium meetings and we talk about sharing and all that sort of thing, you know. We are good citizens of the world, so if we develop something we will immediately share it with someone else. But that’s not the case. You don’t even find the publications behind some of these instruments anywhere. It’s given to the FDA, the FDA reads it, approves it, and it’s not published at all. So one of the exceptions was the psoriasis symptom diary by Novartis, because you know, Novartis wanted to publish it. I was at Novartis at that time and I was wondering why we want to do this, because why do we want to give it to somebody else and save them three years of work and lots of money. But anyway. So new measures are not always available publicly.
And there were only a few labelings based on secondary endpoints—now this, again guys, this is job security, okay—because most of the labels are based on primary endpoints. Now of course if you have a drug that relies on patient-reported outcomes as a primary endpoint, everyone will be working overtime and during the weekends and make sure that you get that, otherwise you’ll have no drug. So real innovation from HCOR or Market Access or wherever you guys might be come
from labelings based on secondary endpoints, and there were only seven labelings based on secondary endpoints, that doesn’t rely on the primary endpoint.
Again, going back to the example of psoriasis, the primary endpoint was a clinician-reported outcome, PASI—I don’t know what PASI stands for but that’s what it is, PASI. And the secondary endpoint, one of the secondary endpoints, was the psoriasis diary. The PASI went in the right direction, great, so you got the indication. Could we have got the indications without the patient-reported outcomes? I think so too, because it was a great drug, it works. But the psoriasis symptom diary also backed up the information, it gave the tailwind to the primary endpoint, which led to a label based on a secondary endpoint, and that was all done for the benefit of the US labeling. Nevertheless, there were only very few examples of labelings based on secondary endpoints.
So the final thought. This sounds like the Jerry Springer show, doesn’t it, I mean the final thought. You don’t watch the Jerry Springer show. No no no, you wouldn’t. I work from home so I watch it all the time. Followed by Ellen Show and followed by a whole bunch of people who come after. Anyhow.
Here’s the final thought. There’s the 21st Century Act recently passed, that encourages faster drug approvals based on biomarkers and surrogate endpoints. This is PDUFA VI. So there is a new initiative—this is what it all boils down to—there is a new initiative to register drugs based on surrogate endpoints and biomarkers. So we don’t wait for all this late endpoints, and endpoints based on patient-reported outcomes and so on. That’s what the 21st Century Act encourages. Whereas the Prescription Drug User Fee Act, the PDUFA V that is currently in operation, encourages inclusion of PRO-related endpoints to drug development. So we have two opposing forces, if you like. So is there a contradiction? I don’t have an answer to this. Both were discussed at the last ISPOR, and I was sort of listening to both discussions and both conversations, half asleep I guess, because this contradiction didn’t click to me until like three days later, when I was working on these slides. So it’s well worth thinking about and maybe discussing later on today, before we have a drink.