The clinical value of microbiologicallaboratory investigations

Summary A sample of 1608 specimens submitted to five microbiological laboratories was identified.Clinical staff on the team responsible for the request were questioned about reasons for the test, and expectations of the results. After the laboratory had reported on the investigation, more senior staff were asked about the contribution of the test result to total patient care. Two in five tests were requested on patients who had no signs or symptoms of infection, themajority (representing 35 per cent of the total) as part of non-discretionary or routine practices. Clinicians expected potential pathogens to be grown in one-third of the tests but were only prepared to name the expected organism or group of organisms in fewer than half of these. Clinicians were able to predict ‘negative’ reports correctly in 62 per cent of cases but predicted ‘positives’ correctly in only 28 per cent. It was intended to repeat 59 per cent of the tests and most of these repeats were to be done as ‘unit policy to screen’ routines. One in five tests were said to have made no contribution to decisions about diagnosis, treatment, management or in any other way. The implications of these findings are discussed with special emphasis on the need to review ‘unit policy to screen’ tests. Other approaches to the reduction in the number of tests which are unlikely to be of value in patient care are also explored.