Adverse event reporting in adult intensive care units and the impact of a multifaceted intervention on drug-related adverse events

Abstract

Background: Adverse events (AEs) frequently occur in intensive care units (ICUs) and affect negatively patient outcomes. Targeted improvement strategies for patient safety are difficult to evaluate because of the intrinsic limitations of reporting crude AE rates. Single interventions influence positively the quality of care, but a multifaceted approach has been tested only in selected cases. The present study was designed to evaluate the rate, types, and contributing factors of emerging AEs and test the hypothesis that a multifaceted intervention on medication might reduce drug-related AEs.

Methods: This is a prospective, multicenter, before-and-after study of adult patients admitted to four ICUs during a 24-month period. Voluntary, anonymous, self-reporting of AEs was performed using a detailed, locally designed questionnaire. The temporal impact of a multifaceted implementation strategy to reduce drug-related AEs was evaluated using the risk-index scores methodology.

Results: A total of 2,047 AEs were reported (32 events per 100 ICU patient admissions and 117.4 events per 1,000 ICU patient days) from 6,404 patients, totaling 17,434 patient days. Nurses submitted the majority of questionnaires (n = 1,781, 87%). AEs were eye-witnessed in 49% (n = 1,003) of cases and occurred preferentially during an elective procedure (n = 1,597, 78%) and on morning shifts (n = 1,003, 49%), with a peak rate occurring around 10 a.m. Drug-related AEs were the most prevalent (n = 984, 48%), mainly as a consequence of incorrect prescriptions. Poor communication among caregivers (n = 776) and noncompliance with internal guidelines (n = 525) were the most prevalent contributing factors for AE occurrence. The majority of AEs (n = 1155, 56.4%) was associated with minimal, temporary harm. Risk-index scores for drug-related AEs decreased from 10.01 ± 2.7 to 8.72 ± 3.52 (absolute risk difference 1.29; 95% confidence interval, 0.88-1.7; p < 0.01) following the introduction of the intervention.

Conclusions: AEs occurred in the ICU with a typical diurnal frequency distribution. Medication-related AEs were the most prevalent. By applying the risk-index scores methodology, we were able to demonstrate that our multifaceted implementation strategy focused on medication-related adverse events allowed to decrease drug related incidents.

Figure 1

Study design. Staff training in adverse events (AE) reporting during 2 months before the pre-implementation period (May 1, 2004 to April 30, 2005). The multifaceted intervention on drug-related AEs was arranged 6 months before its definitive introduction into clinical practice (May 1, 2005). The three structured meetings with the caring staff took place during the same week in the four ICUs.

Figure 2

Distribution of adverse events (AEs) during the day. Most reported AEs occurred between 08:00 and 12:00 a.m., with a peak around 10:00 am. Horizontal dashed line (at 1.8145 AEs/hour) indicates the limit of statistical significant among comparisons of means AEs per hour. SEM, standard error of the mean.

Figure 3

Mean monthly reporting rates of adverse events (AEs) during the entire study period (24 months). Vertical dashed lines indicate the occurrence of structured meetings with the care staff. After each meeting, an increase in AE reporting occurred.