Aim: This paper is a report of a study conducted to identify and test the effectiveness of learning mechanisms applied by the nursing staff of hospital wards as a means of limiting medication administration errors.
Background: Since the influential report ;To Err Is Human', research has emphasized the role of team learning in reducing medication administration errors. Nevertheless, little is known about the mechanisms underlying team learning.
Method: Thirty-two hospital wards were randomly recruited. Data were collected during 2006 in Israel by a multi-method (observations, interviews and administrative data), multi-source (head nurses, bedside nurses) approach. Medication administration error was defined as any deviation from procedures, policies and/or best practices for medication administration, and was identified using semi-structured observations of nurses administering medication. Organizational learning was measured using semi-structured interviews with head nurses, and the previous year's reported medication administration errors were assessed using administrative data.
Results: The interview data revealed four learning mechanism patterns employed in an attempt to learn from medication administration errors: integrated, non-integrated, supervisory and patchy learning. Regression analysis results demonstrated that whereas the integrated pattern of learning mechanisms was associated with decreased errors, the non-integrated pattern was associated with increased errors. Supervisory and patchy learning mechanisms were not associated with errors.
Conclusion: Superior learning mechanisms are those that represent the whole cycle of team learning, are enacted by nurses who administer medications to patients, and emphasize a system approach to data analysis instead of analysis of individual cases.