Measuring medication safety: What works? What doesn't?

By itself, data from hospital incident reports provides an inaccurate and meaningless way to measure the effectiveness and safety of medication use. Research has confirmed that many medication errors go undetected, including those that cause patient harm. Even when detected, errors may not be reported. Practitioners may fail to report errors that have been intercepted before reaching the patient unless the organization clearly defines and communicates all situations that should be reported. Once reported, practitioners may not continue to report similar errors, believing they have adequately informed leadership of the problem. Moreover, failures to remedy reported problems, insufficient feedback about actions taken to prevent further occurrence, complex and time-consuming reporting systems, and fear of personal and professional consequences also inhibit reporting. In fact, focus on error rates derived from spontaneous reporting systems often places undo pressure on practitioners to report fewer errors. However, there are more reliable methods that organizations can employ to measure the safety of medication use and the effectiveness of error prevention strategies.

An understanding that medication safety can be vastly improved by reducing the potential for adverse drug events is key to meaningful measurement of medication use and effective error prevention efforts. While errors that result in serious patient harm occur occasionally, the potential for these catastrophic events frequently lurks in many organizations. As such, measuring potential adverse drug events (PADEs) along with adverse drug events (ADEs) which cause patient harm (minor and serious), provides more meaningful and accurate data upon which organizations can act to prevent errors and show improvement in medication safety. Additionally, error prevention efforts are more effective when measurement is focused on specific high alert drugs or error-prone situations.

For example, PADEs with chemotherapy can be measured by determining the incidence of unsafe order communication (abbreviations; prescribing course doses instead of single doses; failure to indicate mg/m2 dose; failure to round doses over 10 mg, etc.); percent of orders that exceed maximum safe doses during order review; failure to communicate rationale for dose alterations when prescribing; and the availability of current laboratory values (CBC, Cr, etc.), height, weight, BSA and age on order forms. PADEs with administration of drugs to which patients are allergic can be measured by reviewing daily computer reports to determine the incidence of patient profiles without allergy information and the percentage of orders that require intervention to prevent administration of drugs to which patients are allergic. Likewise, reviewing the use of drugs that may indicate an allergic response (e.g. diphenhydramine, corticosteroids) can enhance ADE detection. PADEs with heparin use can be measured by determining the incidence of patients who have not reached therapeutic range within 24 hours of therapy; the incidence of aPTTs over 100 for all patients on heparin; the mean number of aPTTs daily per patient receiving IV heparin; and adherence to heparin protocols. ADEs can be measured by determining the incidence of significant bleeding, administration of PRBC's, or use of protamine.

Continued measurement and visual display of data for PADEs and ADEs together on one chart, with notation at points where various system-based changes have occurred, can clearly demonstrate the effectiveness of error prevention strategies. While spontaneous reports of potential and actual errors can be useful to augment measurement, their solo use in determining a rate is only meaningful when used to demonstrate an increase or decrease in the error reporting rate.