Measuring the Quality of Medication Administration

There are a variety of timing-related factors that impact the quality of medication administration. A drug may be started late, given in dose intervals that do not match the ordered frequency, given despite expiration of the original order, or doses may be omitted altogether.

The criteria for errors in medication timings were summarized by Barker & Allan¹ in 1990 and they remained the standard. Barker & Allan suggested that wrong time errors were defined by deviations from hospital policy. Sixty minutes before or after “scheduled” time appeared to be the standard in 9 studies reviewed. Omission errors were defined as a prescribed dose that had not been given by the time the next dose was due.¹ Based on this definition, a drug given once every 24 hours delivered 23 hours and 59 minutes late would not be considered an omission. Whether the medication system was manual or computerized, the exact time of drug administration was deduced based on business rules and was usually not specified by the prescribing physician. As closed-loop administration systems are adopted by hospitals, the therapeutic assumptions built into the scheduling business rules will be increasingly important. Unwritten business rules, which were applied by nurses in manual systems, will be difficult to apply if the rules are not explicitly defined.

In addition, there was the question of setting and changing drug administration schedules. If a patient had a drug ordered every 4 hours and the initial dose was two hours late, subsequent doses might have been given in four hour intervals from that late administration time to avoid an inappropriate peak. Would all the shifted subsequent doses also be considered two hours late? The criteria for wrong time errors, as suggested by current standards, would not recognize or address these anomalies.

A standardized measure of quality is required in order to assess the different types of administration systems whether they are handwritten, computerized or closed-loop. We theorized that a more useful definition of medication administration errors would be one that (1) correlates drug administration with the achievement of therapeutic medication levels in the patient, altering patient prognosis in the short term; (2) guides “off schedule” administrations by their impact on achieving therapeutic drug levels; (3) considers a nurse’s discretionary authority to shift medication schedules based on the previous two factors; and (4) provides nurses focused priorities for “on time” medication administration.

Defining a Standard

We are proposing a quality standard that will have validity across the hybrid of computerized and manual medication administration systems and processes that evolve over the next decade. This standard has three components:

1. The time interval between an ordered start time and the administration of the first dose for “stat” or “now” medication orders.

2. The average and variance of inter-dose intervals when compared to ordered frequencies for “high alert” (http://www.ismp.org/MSAarticles/HighAlertPrint.htm) and time critical medications.

3. The proportion of medication orders where administration occurs after a discontinued order.

The standards are derived from observation data collected at Vanderbilt and expert consensus. The high-alert medication list is published by the Institute for Safe Medication Practices. This list includes drugs with significant potential for harm when used in error. A similar list could be developed for drugs with significant potential for adverse outcome (either patient safety or economic outcomes) when administration is untimely. Such a list could help staff focus on timely delivery of medications when clinically beneficial.

We found a useful quality summary measure of medication omission to be “missing dose days”. It is defined as the number of therapy days with one or more missing doses. For example, in a pilot study of missing dose days for a single drug (Lasix) in ten randomly selected inpatient encounters in May 2003—the missing dose day’s rate was 5.8% (6/104) days.

Discussion

Standards need to reflect delays in receiving therapy that are expected to have a clinical impact. Our proposal may be of assistance in developing such standards. We plan to quantitate these measures in a CPOE environment and track the changes over time as we advance to computerized closed-loop systems. We anticipate being able to determine whether a meaningful change in medication administration quality has occurred.