 |
 |
 |
 |
 |
 |
 |
|
|
|
|||||||||
|
|||||||||||
Assistant Professor, University of Pittsburgh, Center for Pharmacoinformatics and Outcomes Research, Pittsburgh, PA
Associate Professor, Department of Pharmacy Practice, Northeastern University School of Pharmacy; Adjunct Associate Professor, Tufts University School of Medicine; Clinical Pharmacist, Medical Intensive Care Unit, Tufts-New England Medical Center, Boston, MA
Reprints: Dr. Kane-Gill, University of Pittsburgh, Center for Pharmacoinformatics and Outcomes Research, 918 Salk Hall, 3501 Terrace St., Pittsburgh, PA 15261-0001, fax 412/624-1850, KaneSL{at}upmc.edu
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionConclusionsReferences |
|---|
OBJECTIVE: To determine whether ADE/ADR identification, reporting, and evaluation strategies differ between ICU and non-ICU populations and to characterize ADE/ADR reporting strategies in the ICU.
METHODS: A validated survey was mailed to pharmacy directors at 590 randomly selected hospitals in the US having at least one ICU. A reminder was sent one week after the surveys were mailed. Five weeks later, a second survey was mailed to hospitals that did not respond.
RESULTS: The response rate was 22% (132/590); institutions were predominantly community (68.2%), with 199 or fewer (54.5%) operational beds and 19 or fewer (60.6%) ICU beds. ICU types included mixed medical/surgical (62.1%), medical (48.5%), surgical (31.8%), coronary (29.5%), neonatal (22.7%), and cardiothoracic (15.2%). Operational definitions for ADEs and ADRs varied little between ICU and non-ICU areas, as 92.4% of institutions used the same term for both settings. Triggers were used to identify ADE/ADRs hospital-wide (75%) and were usually the same between ICU and non-ICU areas (88.6%). ADE reporting was nearly always voluntary (94.7%), using paper reports (88.6%), phone calls (22.7%), e-mail (12.1%), Intranet (12.1%), Web-based/Internet (10.6%), or PDA (1.5%). Only 22% of hospitals tracked ICU-specific data.
CONCLUSIONS: ADE identification, reporting, and evaluation strategies are similar between ICU and non-ICU areas. Few institutions currently track ICU-specific ADE/ADR data. The institution of ICU-specific ADE detection and prevention strategies may improve the safety of critically ill patients.
Key Words: adverse drug event, adverse drug reaction reporting systems, critical care, intensive care units
Published Online, July 18, 2006. www.theannals.com, DOI 10.1345/aph.1H088
The intensive care unit (ICU) has the highest incidence of ADEs among patient care units in the hospital. Cullen et al.5 observed the combined incidence of preventable and potential ADEs in the ICU to be nearly twice that of non-ICU areas (19 vs 10 ADEs per 1000 patient days, respectively). The authors also reported that the severity of ADEs was greater in ICU than in non-ICU settings, with more life-threatening events occurring in critically ill populations. The higher incidence and greater severity of ADEs in the critically ill is most likely a reflection of the complexity of care in the ICU environment, the high number of medications used in this population, and the increased vulnerability of the acutely ill to drug toxicity.
Most hospitals in the US have mechanisms in place for tracking ADEs. Survey data indicate that ADE tracking systems are typically overseen by hospital pharmacy departments and most ADEs are spontaneously reported6; however, there is a lack of detailed information about current ADE identification, reporting, and evaluation strategies. Understanding the frequency with which institutions have incorporated standardized operational ADE/adverse drug reaction (ADR) definitions,2,7-14 triggers [defined as signals (actions, drugs, laboratory results) that can aid clinicians in identifying ADRs/ADEs],9,15,16 and evaluation tools will facilitate benchmarking between hospitals.
While clinicians are generally aware that a higher frequency of ADEs occurs in the ICU, it is unclear whether the intensity and methods employed for ADE identification, reporting, and assessment differ between ICU and non-ICU populations. The objective of this study was twofold: to determine whether ADE identification, reporting, and evaluation strategies differ between ICU and non-ICU patients and to characterize current ICU ADE reporting strategies in US hospitals.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
After we obtained institutional review board approval, the validated, teleform-formatted survey was mailed to pharmacy directors at 590 hospitals with one or more ICUs. These hospitals were randomly selected in SPSS from the 3431 acute care hospitals in the US. The cover letter included a request to direct the survey to the employee best suited to respond. A reminder was sent one week after surveys were mailed. Returned surveys were scanned into an Excel database, where results were compiled based on a coding system for each institution. Five weeks later, a second survey was mailed to hospitals that did not respond.
Descriptive statistics analysis for the comparison of ICU and non-ICU data was performed using SPSS version 12.0. Frequency distributions were computed for each survey item to describe the variation among hospitals.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
|
OPERATIONAL DEFINITIONS
The variability in operational definitions used to describe ADEs at the
respondent institutions was consistent with the variability that currently
exists in the
literature.17-19
Institutions usually either stratified adverse events as being ADEs or ADRs
(42.4%) or simply defined all adverse events as ADRs (40.2%). Of the ADE
definitions listed in the survey, the most common definition employed by
institutions (Table 2) included
incidents with potential injury related to a drug (52.3%), although it should
be emphasized that most (56%) institutions used more than one definition.
Sixty-seven (50.8%) institutions provided additional ADE definitions; however,
few (6.8%) used ADR definitions not provided on the survey and that were
specific to the institution. The frequency with which published ADR
definitions are used is provided in Table
3. Many (49.2%) institutions applied several different ADR
definitions in their surveillance programs. Use of operational definitions
varied little among ICU and non-ICU areas, with 92.4% of respondents using the
same terminology throughout their institution.
|
|
SURVEILLANCE PROGRAMS
Identification of Events
Triggers are signals that indicate the presence of a potential ADE/ADR and
can be used to aid clinicians in identifying
ADEs/ADRs.15,17,20
Triggers were commonly (75%) used hospital-wide in surveillance programs as a
means to identify ADEs/ADRs. A list of triggers and the frequency of their use
is illustrated in Table 4. Five
institutions reported using triggers that were not listed in the survey. These
additional triggers included near-miss reports, increased international
normalized ratio, pharmacist review of orders and progress notes, use of
epinephrine, and antidiarrheal prescriptions for patients on antibiotics. Most
(88.6%) institutions used the same triggers in the critical care units as the
rest of the hospital.
|
In addition to triggers, the use of institutional surveillance programs relies heavily on voluntary reporting for the identification of ADEs/ADRs, although 29% of institutions reported using additional methods to identify ADEs/ADRs such as chart review, pharmacist medication administration record reviews, and automated computer searches. A rule-based computer system was used to identify ADEs/ADRs in only 10.6% of hospitals.
Method of Reporting
Adverse event data were obtained most commonly (94.7%) by voluntary
reporting. Institutions relied on paper reports (88.6%), phone calls (22.7%),
e-mail (12.1%), Intranet (12.1%), Web-based/Internet (10.6%), or personal
digital assistant (PDA) (1.5%) systems for voluntary reporting. Computer
programs used to assist clinical documentation included MEDMARX, Zynx, STARS
Web, MIDAS, MSM, or proprietary software developed at the institution. Reward
systems were used by 12.1% of institutions to increase reporting frequency and
incorporated incentives such as cafeteria bonuses, pizza parties, thank you
letters, and pay-for-performance.
ICU ADE Reporting
ICU-specific ADE/ADR data were documented in the minority of institutions
(22%; n = 29), and documentation was not more frequent in larger (>400
operational beds) institutions, with 6.8% (9/132) tracking ADEs/ADRs. The
median incidence of ADEs/ADRs in the ICU per year was 11 (range 1-870) from
the 55 institutions that reported rates. This was a small proportion of those
occurring hospital-wide (132; 6-2985) in the 110 hospitals providing this
information. The median number of ADEs/ADRS reported hospital-wide was similar
for those 55 institutions reporting ICU specific data (median 123) compared
with those that did not (median 166). It should be noted, however, that
location of ADE/ADR occurrences was not routinely documented in most
hospitals.
Evaluation of Reports
The initial assessment of reports is usually performed to verify the
occurrence and further assess the impact of the event. Reports were initially
assessed by multiple (55.3%) individuals or one (41.7%) individual who was
usually (81.8%) a pharmacist and less often (7.3%) a nurse. Multiple member
review teams for ADE/ADR evaluation usually consisted of nurses, pharmacists,
pharmacy technicians, physicians, and risk management employees.
Full-time-equivalent (FTE) allocation for institutions with an individual
reviewer ranged from 0 to 1.0 (median 0.3). The FTE allocation for a multiple
member review team ranged from 0 to 9.7 (median 0.7).
Persons performing the initial ADE/ADR assessments may use objective evaluation tools to help with verification and provide consistency in this process. Many (57.6%) hospitals routinely used ADR detection tools such as internally developed scoring systems (25%), the Naranjo criteria (17.4%), or Medwatch forms (15.9%) to assist in the evaluation and confirmation of ADRs. Of the hospitals using an ADR detection tool, only one reported using a different tool for the ICU; this tool was developed internally.
An evaluation of the outcome associated with each event is important. Most (75.8%) institutions evaluated individual ADEs for outcome severity. The tools used to aid in the outcome evaluation were Medwatch forms (15.2%), institution-specific tools (32.6%), and others (4.5%); 11.4% did not know which tool was used and 36.4% of the institutions did not respond. The "other" category included assessments such as MEDMARX and review by a pharmacy and therapeutics committee.
A secondary review of collected reports occurred at 81.1% of institutions. Secondary reviews are typically performed to determine whether systematic changes within the institution are needed to prevent repeat occurrence of the adverse event and develop a safer environment. Secondary reviewers included a pharmacy and therapeutics committee (67.4%), medication safety committee (19.7%), patient safety committees (17.4%), nurse managers (12.9%), administrators (10.6%), ADE/ADR committees (9.8%), attending physicians (5.3%), and others (12.1%). Those listed in the "others" category included pharmacists, quality assurance committees, risk management committees, and managers.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
The current lack of standardized definitions for adverse events in clinical practice has a number of implications for the ICU clinician. It may compromise the ability to evaluate potential ADEs and hinder benchmarking activities between hospital units or different institutions. Differentiation between ADEs and ADRs in the literature is challenging and often causes uncertainty in clinical practice. The primary difference between ADE and ADR definitions is that ADEs denote injury.17,18 It is important for institutions to clearly outline definitions for ADE, ADR, and injury so that data are accurately collected and consistently evaluated.
The list of triggers that was used in our survey was obtained from the literature.9,15-17,20 The current practice of using triggers to detect ADEs/ADRs on a hospital-wide basis and not specifically for the ICU may lead to an underestimation of adverse events in the ICU. Future efforts should focus on identifying triggers that are most commonly associated with ADEs in the ICU that could include the use of diagnostic testing or laboratory abnormalities (in addition to the more routinely used medication triggers like naloxone).
Voluntary ADE/ADR reporting is an essential component of both institutional and postmarketing medication surveillance systems. The challenges for institutions to reach an adequate intensity of monitoring are well established in the literature.22,23 While the frequency of ADEs in the ICU is at least twice that of other non-ICU areas, one would expect more voluntary reports to be generated from critical care units. However, most institutions cannot accurately determine the number of ADEs that occur in their ICU. In the interest of optimizing resources, the ICU would be a good environment for developing programs focused on increasing voluntary reporting. Only a handful of institutions appear to have simplified the voluntary reporting process by implementing Web-based or PDA systems, and few have formally incorporated incentives to boost reporting frequency. A number of systems, including computer surveillance, are recommended for identifying ADEs/ADRs, although few are commonly employed in clinical practice.9,24
The ADE/ADR review process in institutions usually involves an initial review by a pharmacist and then a secondary review, most often by the pharmacy and therapeutics committee. ADR detection tools are often used to assist in the evaluation of the ADE/ADR reports. Many tools are available; however, none has been designed that accounts for the many unique attributes of critically ill populations.8,12,25 Only one institution reported using an ICU-specific evaluation tool, and this tool had been developed internally. Our results indicate that the most commonly used identification tools are internally developed scoring systems. Of available published instruments, the Naranjo criteria are the most frequently used.25 While the Naranjo scale has been found to be valid and reliable in a small sample of ADRs obtained from the literature, its validity and reliability in the critically ill are questioned.26 An ADE identification tool specifically designed for the ICU is needed to assist in the determination of ADEs in this population.
Quality improvement initiatives should occur as a result of ADE surveillance. Evaluating outcomes provides an understanding of the seriousness of events and should facilitate plans to appropriately allocate resources for ADE prevention. Some institutions lack a uniform approach for evaluating patient outcomes and do not routinely perform evaluations when an ADE/ADR is confirmed. The ADE/ADR surveillance programs should differentiate between ICU and non-ICU events so that specific systematic improvements can be designed for each environment. An effective ADE program should incorporate both an ICU-specific ADE detection and prevention component. Prevention strategies must include markers that identify the event before it results in injury19; therefore, surveillance methods that extend beyond voluntary reporting should be employed.
The presence of board-certified critical care physicians improves outcomes for ICU patients.27,28 Likewise, pharmacist participation in physician-led multidisciplinary patient care rounds improves clinical and economic outcomes and reduces the occurrence of ADEs in the ICU.29,30 Our findings indicate that these clinicians did not provide critical care services at most surveyed institutions. We suggest that the routine inclusion of board-certified critical care physicians and pharmacists in the multidisciplinary team caring for critically ill patients has potential for improving patient safety outcomes.
Although this survey is representative of 38 states, there are limitations to national generalization since some states had only one institution respond. The inherent limitation of any survey with a 22% response rate is the possibility that nonresponders may have different results than responders. Unfortunately, we did not have the hospital demographics of the nonresponders to make comparisons with responders. Our response rate is comparable to that of other national surveys obtaining information on critical care patients.31,32 If the institutions having a developed surveillance system were the more likely responders, then the incidence of ICU-specific reporting may be even lower than the results suggest. The results of this survey also highlight a number of potential areas for research including the identification of more practical approaches for the prevention of ADEs/ADRs in the ICU, the completion of a root-cause analysis to further understand the differences in ADEs/ADRs between ICU and non-ICU environments, the development of reliable and valid tools for the determination of ADEs/ADRs specific to the critical care environment, and the comparison of the impact on patient safety of using various surveillance procedures between ICUs and non-ICUs.
|
Conclusions |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
|
Footnotes |
|---|
This study was presented at the 34th Society of Critical Care Medicine Critical Care Congress, Phoenix, AZ, January 2005.
We thank Michael Bentley PharmD, Peter Herout PharmD, Henry Cohen PharmD, and Jeff Barletta PharmD FCCM for their contributions in developing the survey; Cheri Hill and Michael Keyes for their administrative assistance with managing the surveys; and Teresa McKaveney for her critical review of the manuscript.
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
 |
 |
 |
 |
