Bad Pharma

Why Medical Evidence Goes Wrong

Modern medicine depends on a simple promise. Doctors are supposed to choose treatments based on fair tests, clear results, and honest reporting. But that promise often breaks down long before a prescription reaches a patient. The problem is not usually a single villain or one dramatic fraud. It is a system that allows weak evidence, hidden data, and commercial pressure to shape everyday medical decisions.

One of the biggest problems is missing information. Drug companies can run many trials on the same medicine, then publish the ones that look good and leave the disappointing ones out of sight. When this happens, doctors, researchers, and patients see only a flattering version of the evidence. A treatment can appear more effective and safer than it really is simply because the full record is hidden.

This distortion matters because doctors are trying to make careful decisions in good faith. Ben Goldacre describes prescribing the antidepressant reboxetine after reading published studies that suggested it worked well. Later, it became clear that most of the trial evidence had been withheld. The unseen data showed the drug performed poorly and caused more side effects than other options. Following the available evidence had still led to the wrong choice because the evidence itself was incomplete.

The same pattern appears across medicine. Reviews of antidepressant trials submitted to regulators showed a roughly even split between positive and negative results. Yet once those studies reached medical journals, the published record looked overwhelmingly positive. This did not happen because negative studies were always rejected by journals. More often, they were never submitted at all, leaving the medical literature badly distorted.

The effects are not abstract. Hidden results can lead to direct harm. In one case, an earlier failed study that might have warned researchers about danger in human volunteers was never published, and a later trial ended in severe injuries. In another case, warning signs about heart drugs were not shared in time, and similar medicines were later linked to large numbers of avoidable deaths. When bad news disappears, patients pay the price.

Researchers have tools to deal with uncertainty, especially systematic reviews and meta-analyses, which combine all available trials to reach a more reliable answer. These methods are powerful, but they only work if all the evidence is available. If the missing studies are mostly negative, then even a careful review can produce a false result. A neat statistical answer is still wrong if the underlying data has been filtered.

Attempts to fix this through trial registries and reporting rules have helped, but the rules are full of gaps and often poorly enforced. Older drugs, which make up most prescriptions, are often left out. Regulators also hold large amounts of detailed trial data but may keep it secret under claims of commercial confidentiality or patient privacy. As long as important evidence stays hidden, medicine remains vulnerable to decisions built on half the truth.