But they stress that, although better studies are urgently needed, isolation measures should continue until further research establishes otherwise.

The level of hospital acquired MRSA continues to rise globally. National guidelines in many countries recommend patient isolation to control its spread, but the effectiveness of these measures has never been tested scientifically.

The team analysed 46 studies on MRSA isolation policies published between 1966 and 2000. No conclusions could be drawn in a third of studies, while major weaknesses in others means that plausible alternative explanations for reducing MRSA cannot be excluded, say the authors.

Six studies provided stronger evidence. Four showed that concerted efforts including isolation measures can substantially reduce the spread of MRSA, even when endemic. Yet, in two others, isolation wards failed to prevent endemic MRSA. "These studies indicate a need to investigate precisely how such isolation measures should be used," they add.

Despite all the limitations of existing studies, a lack of evidence of an effect should not be mistaken for evidence of lack of effect, say the authors. "Having considered the evidence, we believe isolation measures recommended in national guidelines should continue until further research establishes otherwise."

"We do need better studies, but we must have faith in the strength of common sense and practical experience when it comes to evaluating infection control measures," adds Dr Andreas Voss, an expert from The Netherlands, where MRSA levels are low.

The fact that The Netherlands and Scandinavia are still able to control MRSA by employing the most intensive "search and destroy" measures should not be ignored. The argument that efforts to control MRSA are insufficient and costly could not be further from the truth, he concludes.