Evil big pharma: myth or material?
pharmafile | April 6, 2011 | Feature | Research and Development, Sales and Marketing |Â Â Les Rose, industry reputationÂ
In the US the False Claims Act regulates those who attempt to defraud the government. In the last 20 years, there have been 165 settlements with pharma companies under the Act, and penalties totalling $14.8 billion have been levied.
The Act was originated in 1863 with the defence industry as its target, but has come into its own when dealing with drug manufacturers.
Indeed, this has built up rapidly over time – 75% of the penalties have occurred in the last five years. Significantly, over half of the penalties were paid by just four companies, GlaxoSmithKline, Pfizer, Eli Lilly, and Schering-Plough. The pharmaceutical industry is now the leading payer of penalties under the Act.
Blowing the whistle on pharma fraud
So what have these companies been doing wrong? The largest proportion of penalties was for promotion of off-label prescribing, a criminal offence. At state level, the next most prevalent abuse was defrauding healthcare providers by over-charging. The next question might be, why has there been this escalation of cases in recent years?
It’s interesting to look at the numbers of cases attributed to ‘whistleblowers’. From 1991 to 2000, these accounted for 9% of the total, but from 2001 to 2010, 67% were sparked off by insider knowledge. My suspicious nature leads me to ask how many cases never came to light in the first of these two decades. Is this rise linked to the fact that whistleblowers can now be rewarded, in some cases extremely generously? It makes sense for the government to pay them $1 million when the company is paying billions. But it might be considered that such people stand little chance of getting another job, so need some financial security.
A global problem?
Well you might say, that’s the US and we all know the private system encourages abuses.
Surely nothing like this happens in other countries? Perhaps not quite like this, but Dr Alain Braillon, lately of University Hospital, Amiens, France, has just reminded me about the benfluorex scandal. The drug, a member of the fenfluramine family, was widely promoted for diabetes and also prescribed off-label for weight loss. It was withdrawn from the market in November 2010 because of cardiac side effects, mainly valvular damage. The closely related drug dexfenfluramine (made by the same company Servier) was withdrawn in 1997, yet the French drug regulatory agency AFSSAPS did not monitor the safety of benfluorex, despite having been publicly warned by doctors working in health insurance agencies in 1998.
It is estimated that 2000 patients might have died from its adverse effects.
In the French case, the opprobrium seems to have been diverted to a large extent away from the company, and towards the politicians.
Government ministers both past and present are staring into the spotlight of unwelcome publicity. Yet one has to ask what the manufacturer’s pharmacovigilance department was making of all this. I searched www.servier.com for benfluorex and its trade name Mediator and came up with nothing.
Dodgy science?
So far my focus has been on sales and marketing abuses. The industry has also come under fire for the way it handles R&D.
In 2009 I attended a debate at The Oxford Union, in which several speakers cited well known cases of bias and distortion. I remember Dr Fiona Godlee, then editor of the British Medical Journal, highlighting that the safety database for Vioxx was founded on nine small studies, and that the safety monitoring board was not free of conflicts of interest. I don’t need to remind you of the eventual fate of that drug.
Mobile goal posts
While it is hard to argue against publications that provide clear evidence of R&D errors and omissions, as well as biasses, these have always jarred with my experience. Yes, I have observed some less than ideal practices from time to time, and more than once I have had to warn a client formally that they were not fully compliant with regulations. But who is? My impression is one of the overall majority of people working their socks off to get things right.
If the general public had the slightest idea of the effort required to put together a submission for a product licence, they probably would not believe it. Why do we have blinded data review? This is where we eyeball the clinical trial data and in particular, ensure that we only have data from eligible patients. We do this before we break the randomisation code, so that there is much less chance of biassing the results. Why do we analyse data from all the patients who had as little as one dose of the study drug, and not just those who complied with the protocol? So that we take the most conservative view possible of the safety of the drug. The safeguards are in place, but they are not perfect because the people using them are not perfect – they are human. So we still have problems, and people still push the envelope of regulation as far as they can.
Pharma good in parts
But recently I heard more encouraging news.
In November last year Dr Adam Jacobs, chief executive of Dianthus Medical, gave a very interesting talk via the excellent website ‘The Pod Delusion’. Coincidentally, he was addressing the same issues as I am doing here, and after I had been engaged to do it. I am quoting some of his talk here with his kind permission.
As I did in an earlier piece in this series, Adam agreed that industry-funded trials do tend to generate positive results for the sponsor’s drug, and like me he explained that the bias might be for innocent reasons. Firstly, industry might fund studies that are better quality and thus detect drug effects more reliably. But this is a fairly speculative explanation, while the matter of publication bias is rather easier to define and quantify. A systematic review from Schott and colleagues published in 2010 found that three of four studies evaluating publication bias found that industry-funded clinical trials were more likely to be published than other trials.
A later study by Bourgeois and colleagues confirmed those findings, and also found that industry-funded trials took longer to be published, possibly because they were usually much larger and took longer to process. However, these trials were more likely to be published eventually. So the accusation that pharmaceutical companies try to bury bad results isn’t really borne out by the data, as overall their publication rate is better than it is for non-commercial trials.
Déjà non-vu
Another common accusation is that of outcome reporting bias. Typically this is where a secondary outcome variable happens to come out more positively than the primary one, so in the report the best outcome is presented as the main finding when it was originally planned to be a minor consideration. I have seen this quite a lot in trials of alternative medicine, especially acupuncture. This question was also addressed by Bourgeois et al., who found that it was statistically significantly less likely in industry-funded trials.
What happens when a published paper is clearly invalid, because of poor practice or fraud? What should happen (and in my view not often enough) is that it is withdrawn by the journal. Actually, it does happen often enough to be quantifiable, and this was done by Woolley and colleagues in a 2009 paper. They found that fraud-based retractions were less likely among industry-funded publications. Much as I should avoid anecdotes, perhaps the most celebrated retraction in recent memory is the paper by Benveniste and colleagues on the effect of ultra-dilute histamine solutions on leukocyte degranulation – claimed to be evidence for homeopathy. The study was repeated under strictly controlled conditions and the paper withdrawn by the editor of Nature, Sir John Maddox.
However, although presented as non-commercial, it emerged that one of the authors was a director of the French homeopathy company Boiron, so whether it falls into the non-commercial camp is open to debate.
Innocent bias
Another possibility of course is that publication bias arises simply because the companies have a pretty good idea of which trials to do in the first place. They are not going to do trials that they know have a poor chance of success for their product.
I don’t think there is anything underhand about this. Every scientific investigation starts with a hypothesis, which is based on a plausible mechanism or some prior knowledge that there might be something going on that needs to be tested. Having said that, of course some trials are conducted in the knowledge that no positive results are likely to emerge.
Some practices are so popular in healthcare, despite a lack of evidence, that studies do have to be conducted to lay the ghost of spurious effectiveness. In such cases, the justification for the study is the very popularity of the practice.
But those studies won’t be carried out by pharma companies, for obvious reasons, so maybe this is part of the bias we observe.
There is also the matter of the business case. The companies will not (or should not) fund trials that don’t have a reasonable chance of adding value to the portfolio.
Perhaps I see a pattern emerging here. From the selection of evidence I have provided, the biggest abuses seem to be in the commercial rather than the scientific areas of the industry.
Whether that is a valid conclusion or not, the effect of any abuse is an adverse one on patients. Ignoring the scientific evidence on safety is glaringly dangerous, yet does still happen, while promoting the drug for unlicensed indications exposes patients to unknown risks.
These are very important messages, but the balanced view is that we are not quite as black as we are painted.
Les Rose is a freelance clinical scientist and medical writer. www.pharmavision-consulting.co.uk
Related Content
Drug development contracts: can lessons be learned from healthcare delivery?
There is a very good question to ask at the beginning of every project: “What …
Medical publishing: can we rely on it?
The explosive proliferation of journals suggests that medical publishing is an attractive business. But it’s …
The numbers game: getting quantitative information to patients
The great physicist Lord Kelvin said: “If you can measure something, and express it in …
