Nov 15, 2005 | How does Novartis do it?
That question resonates throughout the industry. At a time when other pharmaceutical giants struggle to justify pediatric uses of their medicines, or to defend themselves against thousands of lawsuits, Novartis has one of the safest and strongest pipelines in the industry. (See “The Sweet Hereafter,” July 2004 Bio•IT World, page 36.)
Since 2000, Novartis has seen an impressive 13 of its new medical entities (NMEs) approved by the FDA — a figure that is double the numbers of Eli Lilly and Merck combined. Next year, Novartis expects to market new drugs for diabetes and hypertension.
In a recent trial of a new multiple sclerosis drug, FTY720, more than 80 percent of the patients were free from lesions showing active inflammation on magnetic resonance imaging (MRI). That compound started off as a drug for transplantation, where it may yet prove itself, but Novartis scientists had the latitude (and funding) to investigate a second indication.
Novartis is a bit different. It focuses on significantly fewer therapeutic areas than most of its rivals — eight. They are neuroscience, cardiovascular, arthritis, infectious disease, respiratory, oncology, transplantation, and ophthalmology.
That Novartis is the only major company using electronic data capture (EDC) in 100 percent of its clinical trials is a source of bitterness and envy in other companies, which continue to prefer error-prone paper forms. (See “Master of the EDC Universe,” August 2004 Bio•IT World, page 26.) That Novartis is the apparently successful product of a 1996 merger between Sandoz and Ciba-Geigy may also be irksome: A few Big Pharma mergers seem to have produced obese conglomerates — and some unmerged organizations that would like to merge cannot find dance partners.
Novartis’ performance is all the more striking when considering that Novartis’ R&D budget is relatively frugal, ranking fourth or even fifth on most tabulations of scientific budgets in the industry. Novartis sometimes calculates its own return on its R&D investment in a manner that a child could understand. It divides the dollars spent by the number of NMEs approved. The rest of the industry, it seems, spends $3 billion to get one NME on the market. Novartis spends $1 billion.
Those numbers may make some observers bristle, as they are higher than the accepted Tufts University estimates based on the average costs of individually cherry-picked research projects. However controversial they may be, the Novartis calculations do provide a rough approximation of what an actual R&D dollar yields.
| ||Joerg Reinhardt|
For a look at the company’s latest news and preferred technologies, Bio•IT World’s
Mark D. Uehling spoke to Joerg Reinhardt, global head of development at Novartis Pharma and a member of the company’s Pharma Executive Committee (PEC). Reinhardt also chairs the board of the Genomics Institute of the Novartis Foundation. Based in Basel, Switzerland, he is involved in go/no-go decisions at two crucial stages of each project’s life, when it moves out of Phases I and II.
Q. It’s disheartening that you calculate that the rest of the industry spends $3 billion to find an NME. Why is Novartis spending less?
A. What I don’t have is the magic principle that we do, that we follow and nobody else does. That does not exist. But I think what we have been doing since the merger is we did systematically increase the hurdles that our compounds have to face to get to the next phase of development.
How does development work at Novartis?
Whatever is coming from research and enters the development pipeline needs to have shown a biological effect in a decent population and needs to have shown a very good benefit/risk ratio or a good therapeutic window. At many other companies, that is not necessarily the case. [At other companies,] the compound comes from research into the development organization without any data in man. That is not what we are doing.
Did you ever worry that no projects would meet more stringent criteria to proceed?
Initially, that is the concern that people had. On the other hand, the system is adaptive. When people know the hurdle is higher, they come with better compounds.
When Sandoz and Ciba-Geigy merged in 1996, there were 130 active projects. Was it difficult to shrink the portfolio to 75?
It was a big task to cut this portfolio down and focus on the most interesting, the best compounds. That focus did help us to be productive. We did get rid of a lot of compounds that did not have a good justification to be in the portfolio. At the end of the day, [we have to] focus on fewer projects with a higher likelihood of success.
Are 75 projects enough?
When you look at Pfizer or Sanofi or Glaxo, they claim to have more projects overall in the pipeline than we do. I am not concerned about that. I believe that the quality of projects we have, in terms of likelihood of success, is higher than for many other companies. It’s not just a numbers game. It’s also a quality game. Given the fact that our compounds must take this hurdle of proof of concept early on, those that make it into the development portfolio do have a better chance to succeed.
When deciding to move a compound into Phase II, do you think other companies may have more emphasis on marketing or put fewer scientists in the room?
I believe that at that stage, many companies do have stronger commercial interests than we have, and may not have that detailed scientific debate. Very often these meetings have 25, 30, 40 people. But it’s fun. There is real engagement and real discussion.
What happens when considering the transition from Phase II to III?
That decision-making body is called the Innovation Management Board. In that IMB discussion, it is taken for granted that the program that is presented is solid from a scientific perspective, that the data are adequate. But nevertheless, especially [chairman] Daniel [Vasella] has concrete medical questions where he wants to understand more about the data that has been collected or wants to understand what the study design going forward is. We do have technical discussions here as well. The other dimension that is important, especially in the transition from Phase II to Phase III, is, OK, so how will you position this product? What kind of trials will you have, by when, to show what the added benefits are? When will you start the outcomes programs?
What role does technology play? Is it all just process and the quality of the people?
No, obviously, technology is a base prerequisite for success in this industry. On the other hand, I do believe that many big and good companies do use more or less the same kind of technologies and exploit them at more or less the same level. This industry — if you don’t get your technology right, and have highly skilled and great people and good scientists, you don’t get anywhere.
You have a unique approach to EDC at Novartis.
We developed our own system. We did not rely on commercially available systems at that time because it was very expensive and was, when we started three or four years ago, not yet at the level of sophistication that we were really happy with. That is why we started to develop our own system. That helps not only speeding up time between database closure and report writing, but it does also keep costs down. You don’t have all of this expensive paper anymore. You don’t have the endless queries going back and forth before you can finally close the database. That’s a technology where I believe it did contribute significantly to make us productive.
Were there worries about building your own EDC system? Were you involved in that decision?
Yes, I was involved when we debated that technology — four years ago or so. There was not that much technology out there. We had a few trials that we did do with the well-established vendors. But I was complaining about the cost. I said, ‘It’s wonderful, you guys, this technology. But look at what we pay!’ Each trial had to be paid separately. There was no way, at that time, to have an umbrella agreement. You had to pay for each trial and each patient. We decided that was not the way forward. The only other option was to do something ourselves.
In considering all the technologies used in drug development, are there any you’ve been able to unplug? It’s rare to hear of a technology that is no longer needed. Instead, pharmaceutical research technology seems accretive, like a snowball rolling downhill.
This is a growing problem. We, from time to time, go through our IT applications and actually kick stuff out that is obsolete and not needed anymore. If you don’t do that, we ended up with hundreds of applications that needed to be maintained and that did create costs. When you look at who uses what, you find out that there are systems and technologies and applications that are not really needed anymore because they are obsolete. But you still have people who like them, much more than the new stuff. So you need to actively take the toys away, close these applications down, and in a sense mildly encourage people to use new applications. We do this from time to time. Every second or third year, we go through the organization and cut that stuff out. We are now in the process of doing this again.
Can you supply any specific examples of technology that is no longer needed?
Specifics are always difficult. We have over the last three to four years developed specific systems for preclinical data, for clinical data, for safety data to be put together, summarized in a core clinical dossier, an electronic dossier. We have new applications based on data warehouse technology to do that. We have stopped using old word-processing systems.
Has there been resistance?
Not everybody is happy with that. But people using the systems have made their work much more productive. You can share documents with 10 to 15 people at the same time. Which was not the case before where you had to send it from Person A to B to C to D and it took many days and weeks to complete it. Now we can change it immediately. This has taken out a lot of ‘white space,’ as we call it.
White space is development time that is between productive activity and is essentially not needed.
How do you decide when to standardize around a technology companywide, and when to let a particular therapeutic area (TA) choose its own tools?
Especially when you have new technologies, if you try to standardize too early and to impose one standard to all of your scientists, you lose a lot of the potential positive implications of the technology. It’s constantly being developed. We let our people play for quite a while with any new technology. We do give them the freedom from TA to TA to use different standards.
Do you ever insist on a technology?
Obviously, standardization saves costs. But when you talk about implementation of new technology, cost is an important factor but not the overriding one. It’s more to really find out if that technology has positive implications, and if so, in which areas. Too quick, too early standardization limits your options. That’s what I don’t like too much.
Novartis immunologists suspected that FTY720 may have better efficacy in multiple sclerosis than in transplantation — two different scientific fields.
We are quite happy that they did this, that they had the courage to do this. It was out of their primary focus. But nevertheless the data looked so good that they investigated, and it came out positive.
Do you think Novartis drugs have a tendency to have a better safety profile than other companies in the industry?
You can’t separate safety and efficacy. It’s always a matter of these two aspects. Often what you see is that if you squeeze out the last little bit of efficacy, you have the dose so high that you get into potential trouble from a safety perspective. If you want to be safe, you may be at the dose that does not work anymore.
Will the lawsuits over COX-2 inhibitors affect the entire industry?
One reaction that the industry will probably have about the ongoing debate about safety is that the industry is not dosing new compounds to the very end anymore. But [most companies] would rather stay at a dose that is efficacious but still shows a very good safety margin. In the past, that may not have been always the case. That’s a learning that the industry has.
Does Novartis handle this differently?
From our perspective, we pay a lot of attention to safety early on. In the proof-of-concept trial, you cannot fully assess the safety of your product. But you get an idea of what the dose range is to get biological activity. Then you have to assess whether that dose range, from all you know about safety, whether that dose range is, let’s say, high enough or different enough from the potential harmful dose range to justify going on with the compound. It’s only in a few cases where you have such a huge delta between efficacy and potential side effects that you can assume that there would be no risk to the compound. For most of the compounds, there is a certain level of risk left. People will simply have to accept that.
In the future, do you think most drugs will be bundled with diagnostics?
It will take some time to see a broad range of bundled products. The industry is investing in pharmacogenetics more and more. On the other hand, we often see it’s often not a simple yes-no answer. It’s very often not just one mutation or one gene that is of importance; it’s a signature of genes or mutations. It’s a bit more complex than some people might have thought a few years ago. It will take a number of years until we see a broader percentage of bundled products. It will come. But it will take longer than original expectations.
Is there any technology in drug development that has disappointed you, where you wondered if you made the right investment?
One is gene therapy, where the industry for a while did hope that putting a gene in a vector and injecting the vector may be sufficient to cure diseases. Certainly, that is not the case. It’s much more complex than that.
Another technology which did not live up to its expectations is antisense technology. Many companies did invest in antisense for quite a while. That area, however, might see some new interest with the si-RNA technology. But the likelihood [of success] for siRNA will be higher than for antisense because you’re talking about smaller nucleotides, about agents that are easier to deliver than the antisense technology. It will be interesting to see how that new technology is developing.
Is there any technology that has exceeded your expectations?
One technology that has certainly exceeded some of our expectations was monoclonal antibodies. Ten years ago, there were only a handful or very few examples. In the last few years, the success of a few obviously important products, including our own Xonair, for treatment of severe asthma, we have seen that this technology of monoclonals has really shown promise — really great promise. We will intensify our investment in that technology area significantly. You may have seen we did strike a deal with MorphoSys not that long ago to invest much more heavily in human monoclonals. We believe that that technology has gained a lot of momentum over the last few years.
Novartis Discusses Next-Generation Clinical Trials Technologies
Photo by Jurg Waldmeier