Hospital drug safety - role of the pharmacists

Medication errors and adverse drug reactions are frequent in hospitalized patients.
The principle aim of my dissertation was to review the existing data about frequency and risk
factors of these findings and to propose measures for their reduction, focusing on the possibilities
of hospital pharmacists.
In more detail, the aims were:
– To review the literature published between 1990 and 2003 for studies reporting incidences
of medication errors and/or adverse drug effects in hospitals
– To investigate the prevalence of potential drug-drug interactions in ambulatory patients
treated with a statin
– To propose dosage guidelines for patients with liver disease being treated with antineoplastic
drugs. This study was initiated because questions about dose adaptation of antineoplastic
drugs are quite frequent in hospital pharmacies
In the first study, I analyzed the original publications about medication errors and/or adverse
drug reactions in hospitalized patients published between 1990 and 2003, with a focus
on frequency, risk factors and avoidance of problems associated with pharmacotherapy. I
performed a database search (Medline, Embase) for original articles using the terms „medication
error“, „adverse drug reaction“, “adverse event”, „hospital“ and supplemented the articles
retrieved by searching review articles for additional references. The analysis revealed
that medication errors occur with a frequency of approximately 5% of all drug applications,
with a high variability among the 29 studies retrieved. This variability is explained by the way
medication errors are detected (systematic screening of patients or charts vs. spontaneous
reports) and by the way drugs were administered (intravenous drugs have the highest error
frequency). Errors occur along the whole medication process, with application errors accounting
for more than 50% of them. Important risk factors are insufficient pharmacological
knowledge and work overload of the nursing staff, non-computerized transmission of prescriptions
and lack of clinical pharmacists on the wards. Adverse reactions affect approximately
6% of the patients per hospitalization and show a high variability between the 31 studies
retrieved. This variability can be explained by different assessment of the frequency of
adverse drug reactions and by the wards studied. Risk factors for adverse drug reactions include
female sex, age >65 years, polypharmacy and medication errors. These findings allowed
me to propose strategies for reducing medication errors, e.g. to improve the knowledge
about pharmacology of all persons involved in the medication process, computerization
of the entire medication process and the engagement of clinical pharmacists on the wards.
In the second study, we performed a cross-sectional analysis of the prevalence of potentially
serious drug-drug interactions of ambulatory dyslipidemic patients treated with a
statin. Data of patients with dyslipidemia treated with a statin were collected from 242 practitioners
from different parts of Switzerland. The medication was screened for potentially harmful
DDIs with statins or other drugs using an interactive electronic drug interaction program.
We included 2742 ambulatory statin-treated patients (mean age 65.1 ± 11.1 [SD] years;
61.6% males) with 3.2 ± 1.6 (mean±SD) diagnoses and 4.9 ± 2.4 drugs prescribed. Of those,
190 patients (6.9%) had a total of 198 potentially harmful drug-statin interactions. Interacting
drugs were fibrates or nicotinic acid (9.5% of patients with drug-statin interactions), CYP3A4-
inhibitors (70.5%), digoxin (22.6%) or cyclosporine (1.6%). The proportion of patients with a
potential drug-statin interaction was 12.1% for simvastatin, 10.0% for atorvastatin, 3.8% for
fluvastatin, and 0.3% for pravastatin. Additionally, the program identified 393 potentially critical
non-statin DDIs in 288 patients. Our study showed that CYP3A4 inhibitors are the most
frequent cause for potential interactions with statins. As the risk for developing rhabdomyolysis
is increased in patients having drug-statin interactions, clinicians should be aware of the
most frequently observed drug-statin interactions and how these interactions can be avoided.
In the third study, we classified the antineoplastic drugs marketed in Switzerland by the
end of 2003 according to their hepatic extraction in order to predict their kinetic behavior in
patients with liver disease and to give dose recommendations. Dose adaptation for liver disease
is important in patients treated with antineoplastic drugs due to the high prevalence of
impaired liver function in this population and the dose-dependent, frequently serious adverse
effects of these drugs. We therefore classified the antineoplastic drugs marketed in Switzerland
by the end of the year 2004 according to their bioavailability/hepatic extraction in order
to predict their kinetic behavior in patients with decreased liver function. This prediction was
compared with kinetic studies carried out with these drugs in patients with liver disease. Of
the 69 drugs identified, 52 had a predominant extrarenal (in most cases hepatic) metabolism
and/or excretion. For 48 drugs, hepatic extraction could be calculated and/or bioavailability
was available, allowing classification according to hepatic extraction. For 17 drugs, kinetic
studies have been reported in patients with impaired liver function, with the findings generally
resulting in quantitative recommendations for adaptation of the dosage. In particular, recommendations
are precise for 13 drugs excreted by the bile (e.g. doxorubicin and derivatives,
and vinca alkaloids). Validation studies comparing such recommendations with kinetics
and/or dynamics of antineoplastic drugs in patients with decreased liver function have not
been published, however. The study shows that there are currently not enough data for safe
use of antineoplastic drugs in patients with liver disease. We concluded that pharmaceutical
companies should be urged to provide kinetic data (especially hepatic extraction) used for
classification of such drugs and to conduct kinetic studies for drugs with primarily hepatic me
tabolism in patients with impaired liver function allowing to give quantitative advise for dose
adaptation.
The studies show that medication errors and adverse drug reactions are frequent in hospitalized
patients. Medication errors are an important risk factor for avoidable adverse drug
reactions. For two of them, drug-drug interactions and dose adaptation in patients with liver
disease, we performed studies focusing on the incidence and guidelines for their avoidance,
respectively. Hospital pharmacists have an important role both in the prevention and detection
of medication errors and adverse drug reactions.