Evaluation of the role of the critical care pharmacist in identifying and avoiding or minimizing significant drug-drug interactions in medical intensive care patients

Abstract

Study Objective

The aim of the study was to evaluate the impact of prospective review of significant drug-drug interactions (DDIs) occurring in medical intensive care unit (MICU) patients by the critical care pharmacist participating in patient care rounds on improvement of safer and more efficacious medication use.

Study Design

A prospective consecutive 10-week study was conducted in the MICU, St Luke's/Roosevelt Hospital Center (St Luke's site), New York, NY.

This study compared baseline period when clinical pharmacist services were not provided with the period when each patient's profile was reviewed daily during MICU rounds and interactions were minimized. The study examined whether the presence of critical care pharmacist would decrease the number of significant DDIs in the MICU. Impact of decreasing presence of severe DDIs on length of stay (LOS) and discharge status was also evaluated.

Results

Having a pharmacist on rounds resulted in statistically significant decrease in number of clinically important interactions requiring therapy modification, rated D-X (Poisson regression B = −1.036; 95% confidence interval, −1.318 to −0.753; P < .01). The coefficient (−1.036) indicates the incidence rate ratio of 0.35, meaning that the presence of clinical pharmacist in MICU rounds decreased DDI rate by 65%.

According to the multiple linear regression, lower number of DDIs was associated with shorter LOS (P < .01).

Inpatient mortality rate was lower in the intervention group compared with the preintervention group. Number of DDIs was not significantly associated with mortality based on simple regression (P = .45) or multiple regression analysis (P = .09).

Conclusion

Implementing a DDI screening procedure results in significantly lower number of important DDI in the MICU and shortens LOS.

Introduction

Adverse drug reactions (ADRs) are very common in intensive care patients. It is difficult to quantify the true incidence of ADRs, but it can approach 29.7 per 100 admissions, according to some studies [1]. Critically ill patients are commonly prescribed multiple medications that can potentially interact with each other. Drug-drug interactions (DDIs) may be responsible for 3% to 5% of all preventable in-hospital ADRs [2]. Another problem with DDI in the intensive care patients is potential loss of therapeutic efficacy leading to treatment failure. The incidence of DDI leading to treatment failure in medical intensive care unit (MICU) patients is unknown.

Some data documenting value of critical care pharmacist joining MICU team during clinical rounds are available. In one study, a pharmacist on rounds decreased the number of preventable order-writing adverse events from 10.4 to 3.5 per 1000 patient-days (P < .001). There was also a decrease in all ADRs, from 33 to 11.6 for 1000 patient-days (P < .001) [3]. Although in general, pharmacists' participation in clinical rounds was demonstrated to be beneficial, no specific data exist on value of establishing a clinical pharmacy-driven program that would actively screen and prevent DDI occurrence.

Currently, physician order entry system at St Luke's/Roosevelt (SLR) Hospital Center (New York, NY) alerts prescribers of DDIs; however, many interacting medications continue to be administered simultaneously despite these alerts.

The purpose of this study was to prospectively evaluate all MICU patients for the presence of significant DDIs and make recommendations to the prescriber to adjust medication regimen to assure safe and effective medication use. The evaluation and recommendations were done by clinical pharmacist participating in patient care rounds in the MICU. This study compared baseline period when clinical pharmacist services were not provided with the period when each patient's profile was reviewed daily during MICU rounds and interactions minimized. Impact of decreasing presence of severe DDI on length of stay (LOS), discharge status, and mortality was also evaluated.

A prospective consecutive 10-week study conducted between October 16 and December 18, 2008, was carried out in the MICU, St Luke's/Roosevelt Hospital Center (St Luke's site). St Luke's/Roosevelt Hospital Center is a 1076-bed tertiary care teaching hospital. Medical Intensive Care Unit at the St Luke's site is a closed unit that has 12 beds; however, the service census frequently exceeds unit capacity, and surgical/cardiac critical care beds are used to accommodate medical intensive care patients.

All patients admitted to the MICU during weekdays and weekends over the study period were enrolled in the study (intervention group). Clinical pharmacist reviewed medication profile upon admission to identify any significant DDIs.

Patients were observed prospectively on a daily basis by the clinical pharmacist present on rounds Monday to Friday until transfer or death. An interaction analysis was performed each time a new medication was added to the patient's profile.

Lexi-Comp database (Lexi-Comp, Inc, Hudson, Ohio) was used to identify and analyze DDIs. Identified DDIs were confirmed using additional reference sources.

Electronic medication order profile was verified against electronic medication administration record. Patients' age, sex, LOS in MICU, comorbidities, discharge status, and the total number of medications prescribed and administered was recorded.

A total number of moderate and serious interactions (classes C-X; Table 1) was recorded in the DDI data collection sheet. Only serious interactions, such as potentially or definitely requiring therapy modification (classes D-X), were analyzed in detail.

Critical care pharmacist reviewed all potential interactions; only interactions that could potentially result in lack of efficacy or an ADR were communicated to the attending physician or the pulmonary fellow, who had the authority to accept or reject proposed changes.

Data for baseline comparison to document how many clinically significant interactions were not acted upon were collected retrospectively for the months of May to August 2008, when clinical pharmacist services were not available (preintervention group). Pharmacy services were provided by the decentralized dispensing pharmacists with responsibilities to review orders for all critical care units, step-down units, and telemetry floors.

This study was approved by St Luke's/Roosevelt Hospital Center institutional review board on October 10, 2008 (institutional review board no. 08-093).

Baseline characteristics were compared with t test and χ² test, as applicable. Poisson regression was used to examine the effects of pharmacist review and interventions on the number of DDIs. Given that the number of DDIs is a count of rare events and the distribution is highly skewed, we deemed Poisson regression as the most appropriate statistical approach. The following covariates were included: patients' age, sex, number of medications received during the stay, LOS, and number of comorbidities.

To assess the effect of the number of DDIs on LOS, linear regression was applied. Logarithm of LOS was entered into the model considering that distribution of LOS is highly skewed. The following covariates were included: sex, age, and number of comorbidities.

If the effects of the pharmacist intervention on reducing number of severe DDIs were found to be significant, a logistic regression would be carried out to examine whether the hypothesized decrease in DDIs affects discharge status (expired vs alive). Number of medications received, LOS, number of comorbidities, and presence of pharmacist would be controlled for in this analysis.

All statistical analyses were carried using SPSS (SPSS Inc, Chicago, Ill). A P ≤ .05 was considered significant.