Autism’s False Prophets: Bad Science, Risky Medicine, and the Search for a Cure is a new book by Paul Offit. Here he is talking about his book. Enjoy.

There is an excellent paper at the Archives of Internal Medicine describing a seeding trial. The authors of the paper helpfully provide a useful and concise description of a such a trial:

Seeding trials are clinical trials designed by pharmaceutical companies to promote the use of pharmacotherapies that were recently approved or are under review by the U.S. Food and Drug Administration (FDA). Seeding trials are designed to appear as if they answer a scientific question but primarily fulfill marketing objectives. Kessler and colleagues portrayed seeding trials as “attempts to entice doctors to prescribe a new drug being marketed by the company” while the company puts its product in the hands of practicing physicians, hoping that the experience of treating patients with the study drug and a pleasant, even profitable, interaction with the company will result in more loyal physicians who prescribe the drug.

As recently as last week I was being told by one industry employee that seeding trials were a thing of the past, and that industry was no longer interested in them. If they are of the past, then it is a recent past. The Annals of Internal Medicine paper describes the documents, many of them internal Merck memos obtained during legal cases, relating to a seeding trial called ADVANTAGE (Assessment of Differences between Vioxx [rofecoxib] and Naproxen To Ascertain Gastrointestinal Tolerability and Effectiveness). Comparative gastrointestinal tolerability was already being examined with a trial called VIGOR, which was also comparing rofecoxib and naproxen. But, you might say, more data is always useful, so what?

The ADVANTAGE Trial was performed to coincide with the launch of rofecoxib on the US market, and was designed not by scientific staff, not by drug safety staff, but by marketing staff.

In January 1999, before the launch of Vioxx, Merck’s marketing division conceived the ADVANTAGE clinical trial. However, Merck did not reveal the key role of the marketing division and marketing objectives of the study. Instead, physician-investigators, participants, and institutional review board members were told that the purpose of the trial was to measure the gastrointestinal safety of Vioxx. Six hundred investigators enrolled and randomly assigned 2785 patients with osteoarthritis to Vioxx and 2772 patients to naproxen, with a target of 6 participants per site, for a 3-month trial starting on 27 March 1999, approximately 2 months before the drug’s FDA approval on 22 May 1999. A Merck marketing slide set used for the company’s internal purposes stated that a goal of ADVANTAGE was for investigators to “[g]ain experience with Vioxx prior to and during the critical launch phase”.

Review of the Vioxx documents revealed 3 key themes that were related to the design and marketing objectives of the ADVANTAGE trial: The trial emerged from the marketing division with a marketing objective; Merck’s marketing division collected, analyzed, and disseminated both the scientific and the marketing data; and Merck did not reveal the marketing purposes of the trial to participants, physician-investigators, and institutional review board members.

This powerpoint slide is instructive. Orac is appalled:

For science-based medicine to be accepted and trusted by the public, the public must be sure that the clinical trials designed to test the safety and efficacy of new drugs conform to the highest standards of science. More importantly, they must know that these trials are as free from bias and hidden agendas as is humanly possible to achieve. Also, a personal standpoint, I detest Merck’s actions because it will make it that much more difficult for honest clinical investigators like myself to win patients’ trust and persuade them to agree to participate in clinical trials, particularly trials in which a pharmaceutical company is involved but the distrust will spill over to all clinical trials

There’s another important point worth making about this behaviour. Merck withdrew rofecoxib in 2004 because of an increased risk of cardiovascular events. One commentator has argued in the New England Journal of Medicine that the deaths associated with rofecoxib may have run into the tens of thousands, given the millions treated. Interestingly, in their defense Merck state that:

because the literature suggested a hypothetical possibility of both cardioprotective and prothrombotic effects of cyclooxygenase-2 (COX-2) inhibitors, Merck initiated adjudication of cardiovascular events by an external expert panel at the end of 1998 (i.e., before the Vioxx Gastrointestinal Outcomes Research [VIGOR] trial began) for future studies of Merck’s COX-2 inhibitors.

When VIGOR presented in November of 2000, Merck obtained the data in March 2000, it showed an increased risk of myocardial infarction in the rofecoxib group:

The incidence of myocardial infarction was lower among patients in the naproxen group than among those in the rofecoxib group (0.1 percent vs. 0.4 percent; relative risk, 0.2; 95 percent confidence interval, 0.1 to 0.7); the overall mortality rate and the rate of death from cardiovascular causes were similar in the two groups.

So despite Merck already being aware of a potential prothrombotic risk attached to COX-2 inhibitors in 1998 they still continued to fund the ADVANTAGE trial, which did not assess cardiovascular outcomes as a major endpoint. After the publication of VIGOR, which showed an increase in cardiovascular risk, they continued with a redundant trial. In addition, a safety signal was detected from spontaneous reports in October of 2000, as Ralph Edwards, the head of the WHO Monitoring Centre, has stated [PDF]:

This interest in new approaches is understandable in one way because our continuing use of the cheapest globally useful method for safety monitoring ‘spontaneous reporting’ is regarded as a ‘failure’: we ‘missed’ the Vioxx problem. But I must remind sceptics yet again that the signal on Vioxx and myocardial infarction was found using ICSR (individual case safety report) data, and prominently reported by the Netherlands centre (Lareb) at a WHO National Centres Meeting (Tunis, October, 2000) only 6 months after the launch of Vioxx. In my view, any reason for ‘failures’ over Vioxx, or most other product withdrawals, has always been lack of resources in following up signals.

Clinical trials should be planned by scientists. In 2000, there was evidence of a possible cardiovascular signal related to rofecoxib. If Merck, as they admit, were already aware of a theoretical risk of cardiovascular events by 1998, then isn’t it a shame that the efforts that went into making a seeding study to increase the number of people exposed to rofecoxib wasn’t instead put into planning a trial that could assess cardiovascular risks. For that we had to wait until the results of other trials in 2004, after millions of additional people had taken the drug.

Currently, the industry and regulators in the US and Europe are dealing with the new concept of risk management plans [PDF], which are meant to be ongoing documents to monitor the safety of drug and investigate potential risks. It is hoped that such plans in the future might concentrate minds on what is important to the safety of patients. Hopefully, marketing staff will not be involved.

Apologies for the lack of posting here in August. I’ve been traveling for reasons of both business and pleasure. Here are a couple of recent non-blog related items I have written.

Cox AR, Wood KMG, Marriott JF, Ferner RE. Causality and definitions: How general practitioners decide what to report to spontaneous reporting schemes. Pharmacoepidemiology and Drug Safety 2008;17(1):S164 (Presented at the 24th International Conference on Pharmacoepidemiology and Therapeutic Risk Management, Copenhagen, Denmark, 17-20 August 2008)

Cox A, Layton D. Adverse drug reactions: Neuropsychiatric drug reactions associated with varenicline (Champix). The Pharmaceutical Journal 2008;281:129 [link -registration required]

I raise this at the risk of falling out with some bloggers who I respect.

A couple of years ago I had a private discussion with NHS Blog Doc about the issue of confidentiality and blogging. We differed on our views on this. I was concerned that patients were not being asked for consent when case reports, or rather stories, were published on blogs. If you submit a case report to the BMJ or Lancet they will expect you to obtain patient consent. Here’s the Lancet:

Consent must be obtained for all Case Reports and Clinical Pictures.

Here’s the BMJ on their Minerva pictures:

We need written consent from every patient, parent or next of kin, regardless of whether the patient can be identified or not from the picture.

This is not the case in blogging. Stories about patients are routinely used to illustrate posts. It is easy blogging. Such posts are voyeuristic and give readers a vicarious interest if they are not involved in such situations. It’s rubbernecking on the net.

How easy is it to identify a patient? Well, in the case of an anonymous blog, perhaps it would be difficult. But what about those that give geographical details?

Tom Reynolds, author of Random Acts of Reality, is an E.M.T working for the London Ambulance Service, he recently posted a story about a one-year-old boy:

The job is a simple one - pick up patient from their home and take them to hospital as quickly as possible - no thinking required and I don’t even need to do any vital sign measurements on this job.

The patient is a one year old child in liver failure and her parents have just been told that a donor organ may have become available.

When we arrive at the home the whole place is in uproar, it’s late in the evening and every member of the family is scrabbling around gathering things into no small number of bags. Clothes, food and the sort of supplies you need for a very sick little one year old.

I do my best to try and bring a little calm to the chaos but the family aren’t having any of it, they are in near panic and their emotions are somewhere between fear and joy. I know when to admit defeat and I leave them be.

It’s powerful stuff. He writes well. I’ve even seen his blog linked to from an official NHS website.

What if the parents of that boy recognize themselves?

When you go to the doctor, nurse, pharmacist, or are seen by a paramedic, do you really expect your story to appear later on a website for other people to pour over? What right do bloggers have to make use of your story in order to create traffic and readership? What public benefit is being served? In the case of a BMJ or Lancet case report, the report may draw attention to a rare adverse effect of a drug, a new interaction, or a difficult case to diagnose. Even then consent is required. What gives blogs an opt out from standard practice in journals?

It is unfair to pick on Tom Reynolds, and that is not what I intend, especially because a recent paper points out that many medical blogs are guilty of this (via Bad Science):

We identified 271 medical blogs. Over half (56.8%) of blog authors provided sufficient information in text or image to reveal their identities. Individual patients were described in 114 (42.1%) blogs. Patients were portrayed positively in 43 blogs (15.9%) and negatively in 48 blogs (17.7%). Of blogs that described interactions with individual patients, 45 (16.6%) included sufficient information for patients to identify their doctors or themselves. Three blogs showed recognizable photographic images of patients. Healthcare products were promoted, either by images or descriptions, in 31 (11.4%) blogs.

Here’s another example, and another , and there’s lots more out there.

It is quite clearly unacceptable to be the subject of a blog post without consent. There is a difference between health care professionals and others. We should be trusted. We are invited in by patients to share the most intimate of problems. They trust us not to share those problems with others, without their consent.

Professional regulators are behind the curve on this. They must to catch up and quickly. Guidance should be issued and the practice should be stamped out. At the moment blogs are failing patients. They should have the same standard as journals, especially since blogs are publicly accessible in a way most journals are not.

The CTC have been helping a young man with Asperger’s Syndrome with his cycling using their cycling champions programme:

My son Will has Asperger’s Syndrome which for him manifests itself with deep depression and anxiety. Earlier this year at an appointment with his psychotherapist, he was asked to colour in a sheet of paper to reflect his feelings. He coloured the whole page black and used sharp, thick, heavy lines to represent his anxieties but in the bottom corner of the page was a small yellow triangle. The psychotherapist asked “what does this represent?” and he replied “riding my bike”. He went on to explain that when he is out riding his bike everything seems ok and manageable. As a result, we decided that we should ride bikes as a family and, with encouragement from CTC’s Cycle Champions Officer Steve Marsden and affordable second hand bikes bought from Recycle Bikes, we now have at least one long bike ride a week. Will has decided that he would still like to be a scientist, but he now would like to spend one day a week being a professional cyclist.

Adverse drug reactions (ADRs) are a massive burden on the NHS, and a significant cause of morbidity and mortality. When I worked on a cardiology ward, one of the more serious ADRs that occurred was severe haemorrhage associated with streptokinase. Rather than dying of a heart attack, the drug we gave to patients to improve their chances of survival would on occasion be the reason for the demise of the patient due (for example) to a cerebral haemorrhage.

The NPC recently reported on a systematic review of incidence and nature of in-hospital adverse events. The full paper is available as a free-access article at Quality and Safety in Health Care. The NPC note that the study shows:

• The median overall incidence of adverse events was 9.2% with a median percentage preventability of 43.5%.
• The median percentage attributed to medicines was 15.1% (interquartile range 11.9 to 20.4).

The NPC argue that this is in-line with a NPSA report Safety in doses: medication safety incidents in the NHS which they say showed that “between 3.5% and 9% of hospital in-patients experience harm from medicines.” However 15.1% of 9.2% is 1.39%, which is outside the range of the NPSA report. Thankfully, there is another well-conducted study due to be published later this year by the same group that published the 2004 BMJ paper on drug-related admissions to hospital. From what I have seen in presentations, the incidence is more in line with the NPSA report. It is possible that studies which are primarily focused on drug therapy have greater sensitivity to adverse effects of drugs than those which attempt to catch all patient-care events.

Another point worth discussing about the NPC blog post their view on the impact of the study:

The Impact
Adverse events related to medicines are a serious health issue. They cause levels of mortality higher than more high-profile health problems such as breast cancer or AIDS. In addition to the harm caused to individual patients, they also place a significant burden on healthcare resources.

The source from this comment arises from the report To Err is Human, an influential, and often cited, IOM report from 2000. The executive summary uses this comparison:

More people die in a given year as a result of medical errors than from motor vehicle accidents (43,458), breast cancer (42,297), or AIDS (16,516).

I’ve used similar statements myself in order to shock health care professionals into accepting the burden of ADRs as important and worthy of their consideration, but these sort of statements can also feed into some of the more deranged, but not totally uncommon, rants about Big Pharma. Here’s a relatively extreme example from JABS, the UK’s leading anti-vaccination site commonly linked to by the BBC:

big pharma and bad doctors have killed more people than wars in the last 50 years.

For gods sake what more corruption, cover ups and evil does big pharma have to do beside kill a million each year so these sick trolls will stop acting as as shills for them.

Who would you listen to , a mad sick corrupt troll who showns sociopathic tendanicies who thinks its fine for big pharma to commit genocide or real non corporate doctors and scientists who actually care about humanity and big pharma genocide and corruption.

I think sensible discussions about the harms of medicines need to avoid feeding into this meme, and documents like To Err is Human can unfortunately be used by opponents of evidence-based medicine to attack the use of prescribed drugs. What’s missing is the difference between breast cancer and prescribed drugs.

The difference between breast cancer, and the adverse effects of drugs, is that drugs have beneficial effects that, on balance, outweigh the risks they pose to society. For individual drugs when the risks are found to outweigh the benefits, then action is taken such as withdrawal of the drug. There is no corresponding benefit to having AIDS or breast cancer.

Every day individual patients make decisions about risk-benefits when they chose to take a medicine or not. Of course, many could be better informed about that risk-benefit balance, but overall the majority chose to accept the risks for the benefit they will receive. Sometimes patients will accept more risks than regulators think they should. Medicines are more like cars. There is a benefit from the mobility and personal freedom that cars give individuals, which seem to make death in car accidents relatively less important than knife crime, despite the massively higher numbers of deaths in car accidents (I accept the issue of intent makes a difference in this area).

The issue of preventability is also difficult. Some drugs increase the odds of experiencing an adverse event. Not all events will therefore be avoidable. Aspirin increases the risk of a gastric bleed. So not all gastric bleeds could have been avoided if aspirin had not been prescribed.

There is also the issue of monitoring: Coleman et al’s paper in the BJCP neatly explains the lack of evidence for effective methods of monitoring for harms of treatment. Here’s another cardiology example, which I also regularly saw, which Coleman uses to illustrate the difficulties of devising monitoring periods:

The incidence of hyperkalaemia in patients treated with spironolactone for heart failure seems much greater in practice than in the randomized controlled trial that showed the value of the treatment. This seems unlikely to be simply a problem of monitoring. In general terms, patients in clinical practice are sicker and less frequently reviewed than in clinical trials and therefore it is not surprising that in many examples the rate of adverse effects is higher in actual use when compared with the rates seen in the trials. Real patients may also be less likely to adhere to a stringent monitoring scheme, or more likely to have concurrent ill health that increases the chances of finding an abnormal result that is not due to the adverse reaction, or older than trial patients, and perhaps as a consequence more rapidly susceptible to the adverse drug reaction, so that the time between a reaction first being observable and it causing irreversible damage may be shorter. It raises the question of how to devise safe monitoring schemes that will remain effective when treatments move from clinical trials to general use.

So preventability can be in the eye of the beholder, and that eye is often trained on the events using a retrospectoscope, armed with more information, and time, than the clinician making the decision that led to the adverse effect.

And some adverse reactions are not preventable. Take streptokinase, discussed at the start of this post. Even with a perfect diagnosis, and evidence-based treatment, fatal adverse effects can occur. However, on balance streptokinase saved more than it lost, and this should be acknowledged when talking about the harms of medicines.

I am one.

Mobileme sucks in comparison to Gmail.

Photo from iboff.

That is all.

There’s a new website where you can leave a review of your doctor called iwantgreatcare.org. It has already been formally warned by libel lawyers acting on behalf of thirty-seven GPs, of the potential for defamatory comments to be posted by the public. There are some nice comments there too:

Many years ago he literally saved my life and has continued to be the best family doctor I have ever met.

and

Caring understanding, Intelligent and sensitive, Top Professional.

I’m not sure how useful such patient anecdotes are, take this one for example:

He was held in very high regard by the overwhelming majority of his patients. He was also respected by fellow professionals. His patients appear to have regarded him as the best doctor in Hyde. His register was full and there always seems to have been a waiting list. Patients liked him for a variety of reasons. Many would say that he ‘always had time’ for them. His surgeries overran but no one minded because they understood his wish to take whatever time was necessary for each patient. He never hurried them out. He always had time for a few words of a personal nature. Elderly patients and their families were particularly grateful for his willingness to visit at home.

That, you may have guessed, is Harold Shipman.

One of the most interesting comments from the owner of the new rating site is this:

The vast majority of comments have been positive -interestingly, the small number of spoof and malicious comments we have received have been traced back to doctors.

Very strange.

According to some of the more wacko-stalker elements of the UK’s anti-vaccine movement I am part of a government/pharmaceutical industry conspiracy to suppress the self-evident harms of vaccines. Supposedly, Ben Goldacre, Richard Horton and I comprise a supporting act to Roy Meadow in this regard. I am also alleged to be a shill for the industry, presumably paid. An accusation also leveled at me because I am able to use HTML and install Wordpress.

I have to say, this is news to me, flattering though it is to be linked to the editor of the The Lancet, who I do not know, and Ben Goldacre, who I have never met, but know through the internet because of his excellent blog.

So why do these people see conspiracies, where none exist? Charlie Booker explains in relation to the 911-truther movement:

recently I’ve found myself bumping into people - intelligent, level-headed people - who are sincerely prepared to entertain the notion that there might be something in some of the less lurid 9/11 conspiracy theories doing the rounds. They mumble about the “controlled demolition” of WTC 7 (oft referred to as “the third tower”), or posit the notion that the Bush administration knew 9/11 was coming and let it happen anyway. I mean, you never know, right? Right? And did I tell you I’m the Emperor of Pluto?
[...]
Embrace a conspiracy theory and suddenly you’re part of a gang sharing privileged information; your sense of power and dignity rises a smidgen and this troublesome world makes more sense, for a time. You’ve seen through the matrix! At last you’re alive! You ARE the Emperor of Pluto after all!

It should come as no surprise that one of the leading threads in length and numbers of readers in the UK’s leading anti-vaccine organisation’s forum is called “Debunking the Aids virus myth”.