Ben Goldacre's follow-up to Bad Science focuses on how the pharmaceutical industry manipulates doctors and patients. Most people may realize that drug companies design and fund much of the medical research done today (who else would you expect to fund registration trials?), but probably don't know that the results are, if not actively manipulated, disseminated (through publication bias towards positive results) in a way that favors the sponsors. Additionally, once drugs are approved (shown to be safe and more effective than nothing), there's little to no research to show whether they're safer and/or more effective than the current treatments on the market. Finally, there's promotion - drugs are advertised to patients ("medicalizing" issues which may or may not be a problem and inducing them to ask for specific name-brand drugs) and to doctors (who understandably appreciate the convenience industry-sponsored CMEs provide, allowing them to get their credits and cut through the stacks of articles published each month).
As anyone who's read the business section knows, drug companies sometimes do more than selectively release data. Some alter (or actively hide) studies so that ineffective and/or harmful drugs make it to the market. Goldacre argues that the "cure" for this is the same as the "cure" for the less explicit problems. Larger, simpler trials (such as randomly assigning new patients to one of two similar and known to be effective treatments so we can know which one is better), more research from independent (mostly governmental) agencies, an end to drug advertising, and most of all transparency (in funding of studies and conferences, the disclosure of funding to doctors, and the release of the full data from drug studies) would go a long way to improving legitimate research and preventing the more nefarious actions that land pharmaceutical companies on the front page.