Background Numerous studies demonstrate that the Hawthorne effect (behaviour change caused by awareness of being observed) increases health workers' hand hygiene adherence but it is not clear if they are methodologically robust, magnitude of the effect, how long it persists or whether it is the same across clinical settings.Objective Determine rigour of the methods used to assess the Hawthorne effect on hand hygiene, effect size estimation, variations between clinical settings and persistence.Methods Systematic literature review with meta-analysis.Results Nine studies met the criteria for the review. Methodological quality was poor.Data pooling was possible across six studies. The Hawthorne effect ranged from 4.2% to 65.3% with a median of 35.6%. It was 4.2% in one study conducted in intensive care and 16.4% in transplant units. It was most marked when data were collected across an entire hospital and in a group of general hospitals. Differences between wards in the same hospital were apparent. In the two studies where duration was estimated, the Hawthorne effect appeared transient.
Conclusions Despite methodological shortcomings the review indicates clearevidence of a Hawthorne effect on general wards. There is some evidence that it may vary according to clinical specialty and across different wards within the same organisation. The review identifies a need for standardised methodologies to measure the Hawthorne effect in hand hygiene to overcome the dilemma of reporting the potentially inflated rates of adherence obtained through overt audit. Occasional covert audit could give a better estimation of 'real' hand hygiene adherence but its acceptability and feasibility to health workers need to be explored.