Research published today in the BMJ showed that when scientists have financial ties to a company that make a drug the results of trials are more likely to show the drug to be effective. This is particularly worrying because 58% of ‘principal investigators’ had financial links to the drug company. EBP [evidenced base practice] or EBM [evidenced based medicine] involves looking at trials to decide what are the most effective treatments. It is therefore worrying to think that decisions may be made on such skewed data.
The recent scandal regarding Tamiflu is a very costly example. The Government spent £500 million stockpiling Tamiflu [a treatment for flu] in order to be prepared for a Bird-flu epidemic. Yet as far as we know Tamiflu may be no better than a paracetamol at reducing symptoms. Roche the company who makes it refuses to make information available to doctors from the Cochrane centre. The Cochrane centre was set up by Archie Cochrane who served as a doctor in World War 2 and wrote that he believed that much accepted medical practice had no evidence to support their use. He went further and stated that ‘I was afraid that I shortened the lives of some of my friends by their use.’ He promoted evidence based medicine and set up the Cochrane Library of systematic reviews. It has now become the International Cochrane collaboration and involves over 100 countries [see my book Medical School Interviews The Knowledge P69 -71 for further information on Cochrane and Evidence Based Medicine]. Researchers looked at the published trials and began to spot discrepancies. Roche seemed unable to answer the simplest questions, such as how many trials had been conducted. Regulators and agencies around the world appeared to have different information on the same drug. One trial was only published ten years after it was completed. The names of those conducting trials seemed to change arbitrarily. There were also concerns about the design of the trials. Some summaries said patients received a dummy placebo sugar pill, when in fact those pills contained an active ingredient. Other research was done on specially selected patients who would be more likely to recover from flu when given Tamiflu than other patients who were given a placebo. But this is no isolated case. Overall, about half of all clinical trials never get published. What’s more, trials with positive results are twice as likely to be shared as trials with negative ones.
The same is true of trials of other treatments such as certain surgical treatments – for example does removing lymph nodes as well as a melanoma reduce the recurrence of melanoma cancer? Lymph node removal can lead to significant swelling of the arms and may not be beneficial. The Halstead operation for breast cancer was one of the great medical scandals. Halsted developed the radical mastectomy in the 1890s; this procedure removed the breast, skin, nipple, areola, pectoral muscles, and all the axillary lymph nodes on the same side. Even more radical procedures were sometimes used, removing part of the breastbone and ribs to get the internal mammary nodes. These operations would lead to permanent pain, swelling and disfigurement. Trials began to show that recurrence rates were no greater for more limited operations. More was not better! Despite this most surgeons in the USA continued to use the Halstead operation till 1985. Informed consent relies on a patient receiving accurate information and freely making a decision based on that information. If trial information is partially withheld we do not have that information! It is known that about half of all trials are never published These are often the smaller trials with negative outcomes for a drug that a company wishes to promote. A funnel plot can indicate bias. Most results in a systematic review of multiple trials show a normal distribution. However in a funnel plot showing bias the smaller negative trials ‘are missing’
Many of you will be having medical school interviews in the next few weeks. A common question asked is ‘What would you do to save money in the NHS?’ Well, ensuring all treatments funded by the NHS were evidenced based and preventing over treatment would save money and prevent iatrogenic harm [harm caused by medicine] – a win-win situation!
The BMJ is campaigning and calls on ethics committees, funders [such as the NHS] and institutions not to give approval or host trials or fund subsequent drugs unless all trials are published in full and good time. It is also asking for public support so join in and sign the petition at alltrials.net!