Laboratory lies

A survey of British doctors and researchers shows that the pursuit of science is sometimes rather unscientific. The British Medical Journal reported that 13 percent of UK scientists say they've seen colleagues "inappropriately adjusting, excluding, altering or fabricating data," indicating widespread research fraud in the country. "The BMJ has been told of junior academics being advised to keep concerns to themselves to protect their careers, being bullied into not publishing their findings, or having their contracts terminated when they spoke out," said BMJ Editor in Chief Fiona Godlee when the findings were announced.

Some criticized the BMJ's survey methodology as flawed, but the results are in line with other reports: A survey of U.S. researchers a few years ago found that 9 percent had seen colleagues engage in scientific misconduct, such as misrepresenting research findings or submitting false information in order to win a grant. Those results suggest 2,300 acts of misconduct occur annually in the United States.

Reliable information on the problem is hard to obtain, in part because the definition of misconduct varies from place to place (for instance, not everyone considers it "fraud" for a researcher to conceal a conflict of interest). And when misconduct occurs, those who are aware may look the other way.

A 2009 meta-analysis of multiple surveys determined about 2 percent of scientists have engaged-by their own admission-in fabricating, falsifying, or modifying data at least once. Up to a third have admitted to practices some would call questionable, such as "dropping data points based on a gut feeling."

It's rare for an investigation to lead to disciplinary action and the retraction of research papers, but it does happen: In January, the University of Connecticut unveiled a 60,000-page investigative report charging one of its top cardiovascular scientists, Dipak K. Das, with 145 instances of fraud, including cutting and pasting photographs of his research tests.

Much of Das' work touted the health benefits of red wine, and although his papers are widely cited, 26 of them may need to be retracted. Das, who is East Indian, claims the allegedly doctored tests were conducted by students or subordinates, and says ethnic prejudice has fueled the university's charges against him.

Untreatable TB?

Several doctors in India reported late last year that a dozen patients had contracted a strain of tuberculosis bacteria resistant to all medicine. In January the Indian government denied the claim-saying the cases hadn't been clinically proved to be incurable-but if true, it would make India the third country to spawn so-called "totally drug-resistant tuberculosis," after Italy and Iran reported cases a few years ago. According to the World Health Organization, tuberculosis infections were responsible for 1.4 million deaths in 2010.

Some researchers say resistant bacteria have emerged in part because the pharmaceutical industry has been passive about developing new tuberculosis drugs for the past half century, due to lack of financial incentive. But doctors might be able to repurpose some old drugs for the new tuberculosis fight: Leonard Amaral, a bacteria expert at the Universidade Nova de Lisboa in Lisbon, Portugal, told me by email that thioridazine, an antipsychotic drug, would be effective against the Indian strains.