JMIR Publications

ABSTRACT

Background:

There is a paucity of quantitative evidence in the current literature on the incidence of wrong-medication and wrong-dose administration of intravenous medications by clinicians. The difficulties of obtaining reliable data in this area are related to the fact that at this stage of the medication administration chain detection of error becomes extremely difficult. Previous studies have relied on self-reporting of near-miss errors and recognised errors, simulation laboratory studies, chart reviews and observational studies. All of these approaches have inherent weaknesses. Smart pump medication library logs, and their reporting software, record medication and dose selections made by users, as well as cancellations of selections and the time between these actions. Analysis of this data from a facility over an extended time period supplied information that has implications for infusion pump design, medication library software design, and risk management pharmacy strategies. The results can also be applied to strengthen the metrics currently being used for cost-effectiveness analyses for systems such as bidirectional IV Pump-EHR auto-population utilising Bar Code Medication Administration, that are designed to avoid wrong patient-wrong medication and wrong-dose administration errors.

Objective:

To establish, in a reproducible and reliable study, baseline data to show how metrics on the set-up and initial programming phase of intravenous medication administration can be produced from review of medication library near-miss error reports from infusion pumps used in varied disciplines and care areas across a facility in the Middle East. Of particular interest was analysis of user initiated corrections of the more common ‘death-by-decimal point’ errors of incorrect dose/concentration selection, and of wrong medication selection, which is often related to medication-name lookalike-soundalike issues. The study also focused on the length of time it took for clinicians to self-remedy the initial error.

Methods:

A twelve month retrospective review of medication library reports from 2044 wireless-connected modular infusion pumps (846 Syringe Driver Modules, 3662 Large Volume Pump Modules, 62 Patient Controlled Analgesia Modules) used in 15 disciplines/care areas across a facility in the Kingdom of Saudi Arabia was undertaken in order to obtain metrics on the set-up phase of intravenous medication administration. The reporting software used in this study records any attempt by the user to use a dose for individual medications outside of the accepted hospital formulary range. The particular reporter software used in this study also records cancelled infusions and resolutions of all infusion alerts by the user and applies date-time stamps to all of these alerts and actions. Analysis was undertaken using patient anonymised data for infusions in all areas of the facility studied. This included continuous and intermittent infusions, weight based and non-weight based infusions and body surface area based infusions. Decision times of clinicians were calculated from the time-date stamps of the pumps’ logs, all of which are under the control of a central server with universal and accurate time keeping.

Results:

Incorrect medication selections were 3.45% of all medication library alerts and 22.4% of all cancelled infusions. Of these cancelled medications all high-risk medications, oncology medications, and all IV medications delivered to paediatrics and neonates, would require a two-nurse check according to local policy. Wrong dose selection was responsible for 2.93% of all alarms and 19.08% of infusion cancellations. Average error recognition to cancellation and correction time was 27 seconds (22.25) for medication error correction, and 26.52 seconds (24.71) for dose corrections. In terms of the factor of error for dose corrections, generally the potential overdose was not greatly significant (Median 1.5 X the corrected dose), however there were some very large outliers, with 11 corrections of entries with dose error factor of >100 times the corrected dose. The mean average character count of medications corrected from initial lookalike-soundalike selection errors was 13.04, with a heavier distribution towards higher character counts. In terms of the position of the word/phrase error, this was spread between name beginning (69.8%), name middle (21.4%), and name end (8.8%).

Conclusions:

The study identified a significant number of lookalike-soundalike near miss errors, with cancellations of one medication being rapidly followed by the programming of a second. This phenomenon was found to be largely centred on initial mis-readings of the beginning of the medication name, but also showed some incidences of mis-reading in the middle and end portions of medication nomenclature. The value of an infusion pump having the capability to show the entire medication name, complete with TALLman lettering on the interface matching that of medication labelling is supported by these findings. Standardised dosing has commonly been suggested in the literature as being of value in reducing intravenous medication errors; a large capacity medication library is therefore of value in order to provide enough dosing flexibility for clinicians, particularly with weight based infusions, but the study makes it clear that even with fixed doses being available to clinicians the risk of programming errors still exists. Whilst most potential overdoses were not greatly significant, the correction of entries with dose error factors of >100 times the corrected dose indicates the importance of applying concentration limits to all medications that do not have standardised doses available to the user. During the timeframe of this study 603 incidences of inappropriate medication concentrations programming were caught by the medication library’s concentration limits, and whilst the ‘good saves’ seen here were not directly part of the study, (these breaches are captured under ‘reprogram limit alert’ rather than ‘infusion cancelled’) it is clear that IV medication safety at the bedside can be improved significantly by the comprehensive application of a diverse and interrelated group of safety limits. The study also provides a quantitative appraisal of an area that has been resistant to study and measurement; the number of intravenous medication administration errors of wrong medication and wrong dose that occur in clinical areas. This data, when added to self-reporting of near-miss errors and recognised errors, simulation laboratory studies, chart reviews and observational studies more fully triangulates the issue and gives a more accurate idea of the likely value of solutions that offer closed-loop intravenous medication administration. Clinical Trial: Not Applicable