Abstract
We had previously developed and implemented a pharmacy expert system (DoseRanger) that examines drug orders for appropriate single dosage using a commercial rule base and patient specific information. A set of rule adjustments were described and evaluated in order to reduce clinically insignificant alerts. A similar analysis has been performed for daily dose rules provided by the commercial rule base which demonstrated that analogous techniques will be needed.
Introduction
Adverse Drug Events (ADEs) occur in up to 6.5 % of hospitalized patients. Dosage errors in prescribing are one of the most common causes of ADEs. Commercial rule bases can be used to rapidly deploy systems to screen large numbers of orders.
We had previously implemented a commercial rule base (Cerner Multum, Kansas City, MO) with dose range checking rules at a university teaching hospital, two community hospitals, and a home health service. Analysis showed that a small percent of the drug rules constituted a large portion of the alert volume deemed clinically insignificant. We evaluated and successfully deployed the commercial rules by implementing a set of adjustment strategies that could be automatically applied to the commercial rule set2. Here we evaluated the potential to apply similar methods to the rules associated with daily dosing.
Methods
The commercial vendor’s rules for daily dosing in adults were implemented at four institutions in an observation mode for a period of one month. Minimum and maximum daily dose rules based on patient specific information such as age, creatinine clearance, weight, height, gender and BSA were tested.
Results
From 1-24-06 to 2-22-06, 362,166 medication orders were prospectively screened for 752 distinct drugs. Of the 6,280 daily dose rules tested, 6,089 (97%) did not generate any alerts. The remaining 3% produced 11,601 rule violations (3.2 % alert rate).
Conclusions
Implementing a commercial rule set for daily dose rules resulted in a high alert volume. However, a significant number of dosing rules could be implemented without any rule modifications to quickly establish a safety net. The remaining rules may produce substantial numbers of alerts requiring further review by domain experts for possible rule customization. A similar strategy is needed for dose-interval rules to establish a comprehensive drug dosage checking system.
Table 1.
Results
| Facility | |||||
|---|---|---|---|---|---|
| A | B | C | D | All | |
| Screened Orders | 225,932 | 67,461 | 67,304 | 1,469 | 362,116 |
| Drugs Screened | 653 | 526 | 414 | 50 | 752 |
| Drugs Alerted | 264 | 205 | 154 | 3 | 349 |
| Daily Dose Rules Avail. | 6,230 | 6,195 | 6,085 | 6,161 | 6,280 |
| Rules (did not fire) | 5,563 | 5,732 | 5,709 | 6,063 | 6,089 |
| Alerts | 7,101 | 3,175 | 1,320 | 5 | 11, 601 |
| Alert Rate | 3.1% | 4.7% | 2.0% | 0.3% | 3.2% |
Table 2.
Top 5 Alerting Drugs
| Drug Name | Orders Screened | Alerts | Alert Rate (%) |
|---|---|---|---|
| Al hydroxide/Mg hydroxide/simethicone | 3,528 | 1,070 | 30.33% |
| bisacodyl | 3,778 | 999 | 26.44% |
| metoprolol | 4,976 | 705 | 14.17% |
| acetaminophen- oxycodone | 6,255 | 650 | 10.39% |
| ibuprofen | 2,114 | 413 | 19.54% |
References
- 1.Lesar TS, Briceland L, Stein DS. Factors Related to Errors in Medication Prescribing. JAMA. 1997;277:312–7. [PubMed] [Google Scholar]
- 2.Resetar E, Reichley RM, Noirot LA, Doherty JA, Dunagan WC, Bailey TC. AMIA 2005. Strategies for Reducing Nuisance Alerts in a Dose Checking Application. [PMC free article] [PubMed] [Google Scholar]