and audits of prescriptions filled in community pharmacies to monitor dispensing errors (Flynn et al., 2003).




Trained reviewers (nurses, pharmacists)

Most likely to identify events resulting in patient harm; detects more events than self-reports (Jha et al., 1998)

More time-consuming if computer-generated signals are unavailable

Software, trained reviewers

Helps focus reviewer time by using triggers; has the highest positive predictive value for ADEs (see Field et al., 2004); identifies more events than self-reports (Jha et al., 1998)

Best at finding events associated with numbers (Gandhi et al., 2000); availability of electronic data required

Software, trained reviewers

Detects high percentage of ADEs in an efficient manner (see Field et al., 2004)

Electronic record or discharge summaries needed

Providers, report monitoring system and staff

With sufficient data, can identify error and ADE trends; description of event can help trained staff find cause

Detects very small percentage of events

Trained staff to conduct interviews

In addition to advantages of voluntary self-reports, can be performed by attending rounds and nurse shift changes

Detects small percentage of events

Reporting of Medication Errors and ADEs

Voluntary reports, while not appropriate for measuring the actual frequency of errors, are useful as a basis for root-cause analysis and for identification of error trends involving certain medications, doses, forms, and routes. Trend analyses and data mining benefit from having very large databases—hence the efforts being made to increase error reporting and to combine databases (see Chapter 8).

Health care providers can take a number of actions to promote successful medication error reporting in their respective settings. First, they

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement