TABLE C-7 Nursing Home: Administration Errors
Per 100 opportunities/doses—detection method
6 (Cooper et al., 1994)—direct observation
12.2 (Barker et al., 1982)—direct observation
14.7 (Barker et al., 2002)—direct observation
20 (Baldwin, 1992)—direct observation
58 nursing homes identified a mean error rate of 12.2 percent (range 0–59 percent over the 58 nursing homes), where an error was defined as a dose administered or omitted that deviates from the physician’s orders (Barker et al., 1982). The direct observation procedure used in this study detects primarily errors in transcribing and administration. If out-of-date and unsigned orders were excluded, the error rate was 8 percent. The most common error types were unauthorized drug (44.8 percent) and omission (41.5 percent), followed by wrong dose (11 percent), wrong route (2 percent), and wrong form (0.4 percent). Most of the errors involving unauthorized drugs were due to out-of-date orders. Wrong-time errors were not recorded in this study. Because an error is defined as a discrepancy between the drug ordered and the drug received, errors detected by observation may be due to transcription or administration error, but observational studies do not distinguish the phase in which the error originates.
In a 2-year study apparently using observation in one nursing home, Cooper (1987) also concluded that omissions were the most common type of administration error (65 percent of errors). Many of the omissions were caused by patient refusal or sleeping, but the charting often implied that the drug had been administered.
A later study of error rates in skilled nursing facilities and hospitals found an average rate of 21.6 percent in 12 skilled nursing facilities in Georgia and Colorado, using the same direct observation method of error detection and defining an error as a discrepancy between the dose ordered and the dose received. The range of error rates across the 12 nursing facilities was 5.7 to 49.5 percent. The average error rate was not statistically different from the 14.4 percent rate for hospitals (Barker et al., 2002). Excluding wrong-time errors, the rate was 14.7 percent for skilled nursing facilities and 9.9 percent for hospitals. About 7 percent of the errors were judged by a physician panel to be potential ADEs. The rank order of error types was wrong time (9.9 percent of doses, 45.4 percent of errors), omission (7 percent of doses, 32.4 percent of errors), and wrong dose (3.1 percent of doses, 14.2 percent of errors).
Using similar observational methods, Baldwin (1992) detected a 20 percent medication administration error rate in a study of 733 residents of 35 domiciliary homes in North Carolina (error rate range 3–44 percent