Buried amid the seemingly never-ending advance of the U.S. Food and Drug
Administration's (FDA) black-box warnings on the labels of most psychiatric
drugs available today, there is increasing debate on the power of those
warnings to change physicians' prescribing behavior.
Over the last 16 months, the FDA has imposed new black-box warnings on all
antidepressants marketed in the United States and all newer-generation
antipsychotics as well as many older antipsychotic drugs. Last month an FDA
advisory committee, after little discussion or debate, recommended that the
agency add a new black-box warning to the labels of all stimulants used to
treat attention-deficit/hyperactivity disorder (Psychiatric News,
March 3).
However, one issue that was addressed by that committee—the Drug
Safety and Risk Management Advisory Committee— involved speculation
regarding the impact of the black-box warning. Some members hoped the impact
would be significant, akin to the large drop in the number of antidepressant
prescriptions dispensed to children and adolescents that followed warnings
about the potential risk of suicidality. Other committee members were more
skeptical, saying that most physicians today do not read or follow the advice
given in the warning sections of drug labels.
Now, a new study looking at black-box warnings and the extent to which
physicians follow them indicates that 1 in 10 patients whose records were
reviewed was written a prescription for a medication that carried a black-box
warning. More importantly, however, 7 in 1,000 patients received a
prescription for a drug that was contraindicated because of a potential
adverse interaction between the new medication and another drug the patient
was already taking, or because the new drug should not have been prescribed to
a patient with that particular diagnosis.
The study was funded by grants from the Harvard Risk Management Foundation
and Partners HealthCare Information Systems and appeared in the February 13
Annals of Internal Medicine.
Researchers led by Karen Lasser, M.D., M.P.H., an instructor in the
Department of Medicine at Harvard Medical School, reviewed the electronic
health records of patients receiving care from Partners Health Care in the
Boston area to determine how frequently physicians prescribed a drug
contraindicated by a black-box warning and how frequently that prescribing
resulted in harm. Partners Health Care is a group of ambulatory practices
including 40 hospital-based clinics, four community health centers, and seven
community-based practices. All prescribing for these outpatient practices was
done electronically.
At the time of the study, however, the electronic prescribing system
provided physicians with only drug-allergy cross-checking and default dosing
suggestions. The system did not inform prescribers about black-box (or any
other) warnings, check for interactions between the drug prescribed and other
medications the patient was on, check for interactions between the drug
prescribed and laboratory monitoring necessitated, or check for
contraindications for the drug prescribed due to incompatibility with a
patient's diagnosis.
The researchers first identified all prescriptions written in 2002 for any
medication carrying a black-box warning. They then looked at the records of
individual patients who received prescriptions for drugs with black-box
warnings to determine whether the warning should have precluded writing the
prescription for that particular patient. The team designated prescriptions
that should have been precluded as "black-box violations."
Lasser and her colleagues identified 33,778 prescriptions written for drugs
with black-box warnings, representing 10.4 percent of the total of 324,548
outpatient prescriptions written during 2002. However, of those 33,778
prescriptions, only 2,354 prescriptions (0.7 percent of all prescriptions)
were written for outpatients who should not have received the particular
medication prescribed.
The overwhelming majority of those prescriptions —1,744 (74
percent)—were written for medications with black-box warnings requiring
some type of laboratory monitoring that was not adequately completed as
recommended in the warning. A total of 401 prescriptions (17 percent) were
written for patients who had a contraindicated disease, and 209 prescriptions
were written for patients who were already taking a contraindicated drug that
may have resulted in a serious drug-to-drug interaction.
The data compilation for the study was completed in early 2004, prior to
the imposition of the black-box warnings on antidepressants (October 2004) and
antipsychotics (April 2005). In addition, the researchers looked at all drug
classes with black-box warnings in 2002, not just psychotropic medications.
Overall, the researchers found that black-box warnings were most commonly not
followed regarding prescriptions for several antineoplastic drugs and several
anti-fungal medications.
However, prescriptions for several psychotropic medications were identified
to which black-box warnings were frequently disregarded, mostly involving
anticonvulsant/mood stabilizing medications. Nearly 25 percent of patients
(129 of 526) identified by the study as taking carbamazepine had no documented
evidence of monitoring of CBC and platelet counts at baseline and at least
once each year during therapy. Nearly 70 percent of patients (266 of 385)
taking lithium did not have serum lithium levels drawn every two months as
advised in the drug's black-box warning.
Lastly the researchers determined what, if any, impact violating the
black-box warnings had on patients. The researchers found four adverse drug
events (ADEs) resulting from a violation of a black-box warning, three of
which were rated as serious (requiring urgent medical intervention) and one
less significant. There were 92 possible ADEs identified, 18 of which were
rated as having a potential for a fatal or life-threatening ADE, 71 for a
serious ADE, and the remainder for less significant ADEs. In addition, 154
medication errors were identified.
Overall, Lasser and her team found that patients who were female, were aged
75 or older, and took multiple medications (four or more) were at highest risk
of a black-box violation compared with younger male, nonwhite patients, who
took fewer medications.
Lasser and her colleagues said one solution for reducing black-box
violations would be for the FDA to "make these warnings more specific so
they are more readily understandable by providers."
"Adherence to Black-Box Warnings for Prescription Medications
in Outpatients" is posted at<http://archinte.ama-assn.org/cgi/content/short/166/3/338>.▪
Arch Intern Med
2005166338