Rosenbloom_and_Ramsdell _08-34_

Document Sample
Rosenbloom_and_Ramsdell _08-34_ Powered By Docstoc

Personal Digital Assistants and the
Reduction of Medical Error
Mark Rosenbloom, MD, MBA, FACEP
and Edith C. Ramsdell, MD, MPH

S      tudies support that medical error
       is responsible for patient suffering,
       loss of life and billions of dollars
in healthcare costs. New technology, par-
ticularly the “Personal Digital Assistant”
                                               patient injury.”2 Medical error can be clas-
                                               sified into three distinct types, illustrating
                                               the extent to which they affect the every-
                                               day practice of medicine (Table 1).
                                                   Type 1 medical error occurs when the
(PDA), is able to provide accessible point-    provider has established a reasonable plan
of-care medical information. While defini-     that is poorly executed and results in a
tive studies are still necessary, the use of   flawed outcome. Healthcare professionals
PDAs equipped with relevant, reliable and      undergo some of the most extensive edu-
accurate drug and medical references, and      cation and hands-on training of any pro-
calculator software will likely reduce the     fession, yet this kind of error frequently
frequency of medical error.                    occurs. Competent physicians often imple-
   A concise definition of “error” may         ment sound decisions in the wrong manner.
help to understand how the point-of-care       Examples of Type 1 medical error include
use of PDAs can reduce the frequency of        miscalculating a drug dose, overlooking a
medical error. Medical error is defined by     serious “adverse drug event” (ADE) or
the Institute of Medicine (IOM) as the         “adverse drug interaction” (ADI), confus-
“failure of a planned action to be complet-    ing similar drug names, and mechanical
ed as intended or the use of a wrong plan      or incidental occurrences during medical
to achieve an aim.”1 It has been further       procedures or surgery.
defined as “a mistake, inadvertent occur-          Considering the typically overwhelming
rence, or unintended event in healthcare       caseload and stress level of most physi-
delivery which may, or may not, result in      cians, otherwise simple exercises in arith-

Mark Rosenbloom, MD, MBA, FACEP, is an adjunct Associate
Professor of Medicine at Feinberg School of Medicine at
Northwestern University. He is the CEO & Executive Editor of PEPID,
LLC, an electronic medical publishing company. Edith C. Ramsdell,
MD, MPH, is an Assistant Professor of Medicine at Feinberg School
of Medicine, Northwestern University. She is a senior writer and
editor for PEPID, LLC.

36 Harvard Health Policy Review

Medical Error Types     Definitions                                      Causes
Type 1                  Good Plan, Flawed Execution                      Miscalculations
Type 2                  Flawed Plan, Good Execution                      Misinformation
                        1."Classic-Flaw" - Never Correct
                        2."Fashionable Flaw" - Currently Obsolete
Type 3                  Inefficient Plan or Execution                    System Inefficiencies
                        1. "Fear-Flawed"

Table 1. Summary of Medical Errors.
metic, memory, phonetics or hand-eye              of the over 10,000 drugs available can also
coordination may prove problematic with-          prove fatal. Prescription translation error
out additional assistance (from a PDA, for        combined with the increasing number of
example). Drug dosing error often results         similar names such as Celebrex, Cerebyx
from simple miscalculation, including mis-        and Celexa can lead to patient morbidity
taken conversions between Imperial and            and mortality.
Metric units, frequency of weight-based              Procedural or surgical flaws are also
doses and proper drug concentrations. A           Type 1 medical error. Incidental surgi-
single decimal place error can result in a        cal error, mismatched blood types and
ten-fold difference in the calculation of a       wrong site surgery are examples of situ-
drug dosage.                                      ations where highly trained professionals
    Pediatrics is particularly susceptible to     using state-of-the-art equipment establish
Type 1 error. Children’s weights can vary         a good plan and execute it poorly.
greatly from day to day, especially in the           Type 2 medical error occurs when a
neonatal intensive care unit, and most            provider initiates a flawed plan irrespective
children and young infants cannot com-            of execution. There are two sub-categories
municate the nature or extent of their ill-       of Type 2 medical error: “classic-flaw” and
nesses. It is not surprising that the pediat-     “fashionable-flaw.” Because most Type 2
ric population experiences a 300% higher          medical error is a result of faulty plan-
likelihood for dosing errors than adults.3        ning based on incomplete or insufficient
Since doses can vary as much as 400-fold,         knowledge, most can be prevented with
errors in weight-based dosing calculations        improved information and protocols.
may be easily missed.                                The “classic-flaw” is Type 2 error
    Two more examples of Type 1 error             caused by medical practice that is never
are preventable ADI and medication name           correct, yet nonetheless seen within the
confusion. There are tens of thousands of         medical community. Examples of “clas-
chemical constituents in the foods, drugs         sic-flaw” include EKG misinterpretation
and remedies we ingest. The possibility of        resulting in the inappropriate administra-
adverse drug-drug, drug-food and drug-            tion of thrombolytic therapy to a patient
herbal reactions is foreseeable. It is virtual-   complaining of chest pain, or sending a
ly impossible to be aware of all the chemi-       comatose diabetic patient to radiology
cals patients have in their systems and           prior to checking glucose levels.
their possibly lethal interactions or ADEs.          The “fashionable-flaw” is medical er-
Similarly, name confusion between some            ror resulting from fluctuating medical

                                                                    Vol. 5, No. 1, Spring 2004 37

guidelines where practices fall in and out     quent litigation.
of favor. Failing to adapt to practice pat-       Inefficient execution is also a function
terns quickly enough can result in error. It   of cumbersome documentation systems
is important to remember a “fashionable-       operating within healthcare facilities. Inef-
flaw” is a medical practice that was not       fective systems result in the loss of a criti-
always considered error. For example, a        cal element of clinical medicine – time.
physician may use full tidal volumes rather    Physicians have less time to spend with
than “permissive hypercapnia” to treat an      their patients when they are forced from
intubated asthmatic patient. While this        the bedside to complete the large vol-
used to be standard care for such a patient,   ume of documentation necessary to pro-
it would now be considered Type 2 medi-        cess patient records and complete orders.
cal error, because this practice can result    Nurse direct patient time is similarly com-
in over-inflated lungs, which will require a   promised when diagnostics and treatments
protracted ICU recovery if they collapse.      are delayed while paperwork is completed.
    Simple problems should not demand
complex solutions. However, over-worked,
over-litigated physicians often employ an      Prevalence
inefficient “shotgun” approach to treat-
ing patients that leads to unintentional       The IOM report, To Err Is Human, brought
secondary events, known as Type 3 error.       the medical error problem to the fore-
Such extraneous actions involve ordering       front of national debate in 1999.1 This
unnecessary diagnostics that may result        study estimated the prevalence of medi-
in a “false positive.” This leads to addi-     cal error in the U.S. healthcare system. It
tional testing and unnecessary treatment.      reported that 44-98 thousand people die
In a “systems” approach, these errors are      annually (~270/day) in U.S. hospitals as a
a consequence of organizational or situa-      result of error. They further estimated that
tional pressure and are considered “forced     more than half were preventable. This fig-
errors.” An otherwise healthy patient may      ure far exceeds annual mortality rates due
present to the emergency department            to AIDS, vehicular accidents and breast
(ED) with a sprained ankle, only to have       cancer – issues that have clearly elicited
the attending physician initiate treatment     far greater national attention and fund-
for high blood pressure and heart rate. Or,    ing. While some researchers assert that
a physician might order a complete cardiac     the IOM report exaggerated statistics by
work-up for every patient seen with any        not sufficiently accounting for terminally
type of chest pain. Obvious symptoms de-       ill patients, few professionals doubt the
manding attention may be temporarily ig-       severity of the medical error crisis or the
nored to follow time-consuming measures        need for system-wide reform.4
(e.g., insurance pre-certification, non-vi-        Americans are increasingly concerned
tal lab work, physical exam maneuvers).        about a healthcare system that is not able
While most physicians find these practices     to control either costs or preventable er-
medically unnecessary, they admit that this    ror. It has been estimated that additional
procedural inefficiency is the only way        healthcare expenditures, lost income and
they can protect themselves from subse-        disability from preventable medical error

38 Harvard Health Policy Review

costs the U.S. economy $17-29 billion an-       that physicians deviated from accepted
nually.5 These expenses come indirectly         Advanced Cardiac Life Support (ACLS)
out of Americans’ pockets as increased          guidelines in 35.2% of emergency resusci-
medical charges and insurance premiums.         tations meaning one-third of all resuscita-
While medical costs continue to rise, every     tions were flawed based on Type 2 medi-
dollar spent on correcting error is a dollar    cal error.8 Other research indicates 71.2%
not spent on preventative or general medi-      of physicians caring for adults with se-
cal care.                                       vere pneumonia did not follow American
    While a majority of medical encoun-         Thoracic Society (ATS) guidelines in their
ters occur outside the acute care setting,      treatment plans.9 The study concluded that
in places such as nursing homes, surgical       guideline non-compliance increased mor-
centers, physician offices and the home,        tality 450%. Conversely, Suchyta et al. re-
statistics indicate that ambulatory and hos-    ported that protocol adherence decreased
pitalized patients are at equal risk for Type   hospital admissions over 56%, reduced
1 medical error. Over 55% of ambulatory         outpatient antibiotic costs $45 per treat-
care patients are at risk for a preventable     ment and reduced in-house patient cost
ADE.6 Further studies report that more          more than $4,400.10
than 50% of hospitalized children may be            Accepted pediatric guidelines are simi-
subject to medication dosing error.3            larly ignored. Young reported that primary
    Quantifying Type 2 medical error is         care physicians treating febrile neonates
more difficult. Review of the literature        chose treatment strategies differing from
suggests few studies have been conducted        accepted guidelines in 60-89% of cases.11
to determine patient morbidity and mor-         Another study stated that, 92.4% of the
tality from physicians’ ignorance of cur-       time, pediatricians, emergency physicians,
rent “best practices” and protocols. Studies    family practice doctors and nurse prac-
suggest 1) physician lack of awareness, 2)      titioners chose the wrong management
protocol non-compliance and 3) protocol         strategies when presented with a child suf-
obsolescence as the causes of significant       fering from a minor head injury.12
preventable injury.                                 Many protocols are outdated. Shekelle
    Cranney et al. reported many family         et al. suggested that 50% of guidelines are
practice physicians were not aware of the       obsolete within 6 years of development,
existence of evidence-based guidelines.7        and that many are out-of-date within 4
The study further stated that physicians        years.13 This is significant as the typical cy-
reported 1) that available protocols were       cle for textbook updates is 4-5 years, mak-
outdated or 2) limited access to proto-         ing many textbooks obsolete by the time
cols in settings where computers were not       they are published.
available. Cabana et al. support this find-         “Forced error” is prevalent throughout
ing.6 The authors cited lack of awareness,      the entire medical establishment, as statis-
familiarity, time, required resources, proto-   tics show. During the course of ten years,
col accessibility and excessive volume of       the typical emergency room physician will
information as barriers to successful pro-      see an average of 50,000 patients. Odds
tocol use.                                      are 1 in 10,000 will suffer a myocardial in-
     In their study, Cline et al. reported      farction within the subsequent 30 days, ir-

                                                                Vol. 5, No. 1, Spring 2004 39

respective of any diagnostics the attending       Incorporating EBM protocols into well-
physician does or does not perform at the         designed programs is critical for successful
initial encounter.14 The reality that five law-   electronic system utilization.
suits await every ER physician who does               In anesthesiology, the death rate
not fully “work up” every patient with re-        dropped from 1:10,000 to 1:250,000 with
mote chest or cardiac symptoms explains           the implementation of protocols, stan-
why physicians order diagnostics and treat-       dardization of equipment and dependable
ments unnecessarily. Unfortunately, these         procedures.15 In critical care there was a
actions expose patients to Type 3 medical         400% increase in survival of patients with
error that would otherwise be avoided.            severe respiratory distress after the imple-
                                                  mentation of computerized protocols.16
                                                  Additional research has shown that acces-
Managing Medical Error: A                         sible bedside resources effectively reduced
Solution                                          error. Having a pharmacist on adult ICU
                                                  rounds, for example, reduced medication
In practice, the management of medical            order error by as much as 66%.17 Similarly,
error typically begins with “shame and            a pediatric study predicted that ward-based
blame” where physicians are held person-          clinical pharmacists could reduce potential
ally responsible for mistakes. “Bad out-          adverse drug events in children by 94%.18
come” is automatically associated with                To address systemic error, the Joint
malpractice. Ongoing fears of legal re-           Commission on Accreditation of
prisal and public disclosure serve to dis-        Healthcare Organizations (JCAHO) insti-
courage accurate error reporting. As every        tuted widespread policy reform, requiring
patient record is subject to legal discovery,     accredited hospitals to report all adverse
this destructive, finger-pointing approach        events.19 Fear of revocation of hospital ac-
necessitates the creation and utilization of      creditation and the release of all hospital
new risk-management systems. Given the            data to third parties, including the media,
ongoing medical error crisis and its human        will likely preclude effective reform. To be
toll, current obstacles must be addressed         successful and feasible, a systemic solu-
before meaningful advances in reducing            tion must equally address the three types
medical error and promoting patient safe-         of medical error prevalent today in health
ty can be realized.                               care: miscalculations, misinformation and
    A second IOM report discussed the             system inefficiencies.
need for “the efficient and reliable pro-             Programs developed for electronic data-
duction of goods and services according           bases allow for consistent protocol access
to the highly personalized needs of indi-         by clinicians and offer a simple solution
vidual customers.”20 While always under           for protocol non-compliance. If all system
review and subject to change by experts in        providers have access to the same data-
each field as new evidence emerges, evi-          base, “access” theoretically acts as a “reli-
dence-based medicine (EBM) protocols              ability check” to protocol awareness and
already exist for a variety of healthcare ar-     adherence. Electronic storage and transfer
eas including ACLS, ATLS, PALS, APLS,             of medical data also allows for more ef-
BLS, Asthma, JNC7 and ATS guidelines.             ficient and reliable patient management

40 Harvard Health Policy Review

by improving the flow of patient orders,       challenge is to create a true point-of-care
facilitating the exchange of diagnostic re-    medical reference. Point-of-care solutions
sults and simplifying patient cross-cover-     are now available for different specialties
age. These point-of-care communication         that satisfy both the needs of the provider
systems have been shown to reduce pedi-        and the system in which they work without
atric medication error.7                       increasing the management burden: this is
    The challenge is to develop a system       the goal.
that does not secondarily overburden al-            Providers will welcome such proto-
ready overworked and inefficient orga-         cols and technology because the systems
nizations. For the most part, personal         can provide customized access to the best
computer (PC) based technology lacks the       sources of up-to-date information. Hospi-
portability and practicality of a point-of-    tals and other healthcare systems will also
care tool, as the dynamic of the interface     likely welcome the technology because
often requires more information manage-        these systems offer 1) net time savings
ment. Systems not only remain difficult        by decreasing inefficiencies while increas-
to use, but are often incapable of com-        ing productivity and 2) system wide reli-
municating with each other. Further, new       ability checks by decreasing medical error,
Health Insurance Portability and Account-      improving risk management and tracking
ability Act (HIPAA) guidelines and other       protocol use and compliance.
privacy concerns limit data accessibility          It has been reported that physicians,
and transfer. As a result of these second-     including those with pharmacology train-
ary issues, several healthcare systems have    ing, routinely miss over 85% of significant
abandoned electronic solutions.                ADIs identified by a simple drug-interac-
    Consider California’s Cedar-Sinai Hos-     tion program.22 These medication order
pital, where a $34 million system was re-      errors were reduced 83% with an order
cently dismantled.21 System failure was due    entry system capable of providing dosing
to Type 3 medical error that outweighed        information and calculations, and check-
predicted benefits. Doctors complained         ing for allergies and interactions.23 Other
that the system, designed to increase ef-      studies suggest that computerized order
ficiency, actually decreased direct patient    entry can reduce potential ADIs in chil-
care, forcing them to devote too much          dren by 93%.18
time to data entry and system manage-              Ongoing studies further reinforce the
ment. At least six other hospitals have also   premise that well designed and properly
eliminated their paperless systems due to      implemented computer protocols are bet-
end-user resistance, system overburden         ter than humans at rapidly making basic,
and other communication hurdles.21             yet critical, bedside decisions. This is par-
    Medicine has proven that EBM pro-          ticularly true of PDAs, which can access
tocols and technology improve outcomes         volumes of information at the bedside in
and save lives. The challenge is to identify   less than 1 minute.24 Personal Digital As-
the best solution for unique demands with-     sistants prove ideal for these unique point-
in specific healthcare environments while      of-care demands for a number of reasons
delivering these protocols to the practi-      (Table 2). David Siegal, MD, JD, wrote in
tioner in “real-time.” In other words, the     ACEP News: “It is entirely conceivable

                                                              Vol. 5, No. 1, Spring 2004 41

that [the PDA] will become, or has be-         1.   Compact
come, the standard of care.”25                 2.   Fast Processors
    When considering the PDA potential, it     3.   Abudant Inexpensive Memory
                                               4.   Easy Interface with PC
is important to describe how the PDA of-
                                               5.   Excellent Battery Life
fers a solution for reducing the three types   6.   Excellent Readability
of medical error. Point-of-care patient        7.   Wireless Capabilities
management software will significantly de-     8.   Phone/Camera/Video/Audio Capability
crease Type 1 medical error by confirming      Table 2. PDA Features.
current medications, allergies, lab values,
surgical site and other vital information.     “Real-world” scenario
Built-in calculators can verify dosing, drip
rates, lab abnormalities and key medical       A 7-year-old child with a penicillin allergy
parameters. Software can also automati-        presents to the ED with a fever of 103oF
cally check for ADI and provide ADE            and a severe headache. The child has a
information. Such easy access to patient       known history of Tourette’s syndrome
care information will allow better patient     and takes the antipsychotic medication
cross-coverage and should decrease the es-     pimozide to control the symptoms. The
timated 500% increase of risk of an ADE        attending physician suspects a case of
during patient transfer.26                     acute sinusitis.
    A PDA can address both the “clas-              In a non-technological system, the phy-
sic-flaw” and “fashionable-flaw” Type          sician must rely on and trust memory (Type
2 medical errors by suggesting diagnos-        2 medical error), or review a textbook to
tic and therapeutic plans based on EBM         determine the best diagnostic strategy.
guidelines and protocols. Physicians, other    This process is time-consuming (Type 3
health professionals and healthcare sys-       medical error) and the available informa-
tems will be able to confirm orders and        tion may prove outdated or incomplete
protocol use and compliance. Immediate         (Type 2 medical error). The physician may
access to peer-reviewed protocols can help     consult a pediatric ENT specialist, which
change standards of care and eliminate in-     also demands extra time and additional re-
efficient plans based on system overbur-       sources (Type 3 medical error).
den, decreasing the fear of litigation and         Given the child’s penicillin allergy, the
reducing Type 3 medical error.                 physician must know to not use the com-
    To fully understand the impact that the    monly administered amoxicillin (Type
electronic age (specifically PDA technol-      1 medical error), and must review a text
ogy) will have on healthcare systems, the      (Type 2 medical error), or place a call to
practice of medicine and the elimination       a pharmacist to determine a reasonable
of medical error, consider a “real-world”      alternative medication (Type 3 medical er-
scenario using three methods of patient        ror) and possible drug-drug interactions
management: “old school” (non-techno-          (Type 1 medical error). Once an appropri-
logical), an electronic database and the       ate, non-interacting antibiotic is identified,
PDA.                                           multiple calculations to determine dose
                                               and concentration must be performed
                                               (Type 1 medical error). Only after this

42 Harvard Health Policy Review

Patient Specific Programs      Patient Tracker
Medication Information         ePocrates
                               Lexidrug & Interact
                               Davis's Drug Guide
                               Pocket PDR
                               PEPID PDC
Calculators                    MedCalc
                               PEPID MC
Diagnostic & Treatment         Five Minute Clinical Consult Series (5MCC)
                               Harrison's Manual of Medicine
                               Washington Manual
                               Up-to Date
                               PEPID ED
Table 3. Selected Available PDA-Based Software Solutions.

time-consuming, possibly erroneous and         lel access to the same data.
“fear-flawed” scenario, can our physician          In the above scenario, the physician
safely write a prescription and send the       could use a medical information program
child home.                                    (e.g., 5 Minute Emergency Consult) to re-
    Currently, good resources for disease      view disease presentation, diagnostic cri-
and drug information are available on the      teria and appropriate treatment strategies.
Internet (e.g., eMedicine and Up-to-Date)      Then a separate pharmacological refer-
and creditable software (e.g., PDR and         ence (e.g., ePocrates) must be accessed to
LexiDrug) is also available for the PC or      determine appropriate medication and to
laptop. Unfortunately, none are designed       assess potential drug-drug interaction. A
for quick access and instantaneous point-      medical calculator is then needed to com-
of-care decision-support. The inefficient      pute dosing conversions. Finally, a tradi-
use of physician time required to leave a      tional calculator is needed to compute the
patient’s bedside, log onto a PC or Internet   specific patient medication dose. While
site, search for desired information, review   there is less chance for medical error us-
lengthy “text-book” format and decipher        ing non-integrated PDA software than in
inconsistencies creates significant Type 3     the methods described above, the system
as well as potential Type 2 medical error.     is not fluid and leaves room for error, es-
    There are several PDA-based software       pecially Type 3 medical error.
solutions available (Table 3). At present          In an integrated software solution all
the software for PDA systems is either         diagnosis, treatment, medication choice,
integrated or non-integrated (Table 4). In     dose and calculations are nested within a
the non-integrated software, several sepa-     single, seamless program. The program
rate PDA programs must be used in se-          allows for easy access between diagnostic
quence to obtain the needed information,       and pharmacological information, drug-
while integrated software allows for paral-    drug interactions, and medical conversions

                                                              Vol. 5, No. 1, Spring 2004 43

                                 Open Medical Reference         • Go to index
                                 Eg. 5MCC                       • Locate "Sinusitis"
                                                                • Go to diagnostic criteria
                                                                • Go to treatment
                                                                • First line treatment: Amoxicillin (patient allergic)
                                                                • Find alternative: Macrolide-Clarythromycin
                                                                • Close medical reference
       Non-Integrated Software

                                 Open Pharmacological Reference • Go to drug interactions
                                 Eg. ePocrates                  • Enter Clarithromycin
                                                                • Enter other patient medications: Pimozide
                                                                • Note dangerous interaction
                                                                • Consider another Macrolide-Azithromycin
                                                                • No interaction noted
                                                                • Go to drug information
                                                                • Enter Azithromycin
                                                                • Note dosing information
                                                                • Write it down
                                                                • Close the drug reference
                                 Open Medical Calculator        • Find a metric converter
                                 Eg. MedCalc                    • Enter patient's weight in pounds - convert to kilograms
                                                                • Note result
                                                                • Close metric converter
                                 Open Traditional Calculator    • Multiply weight in kilograms by daily dose in milligrams
                                                                • Divide by frequency of administration
                                                                • Divide does by antibiotic concentration
                                                                • Record correct dose
                                 Eg. PEPID                      • Go to index
       Integrated Software

                                                                • Enter "Sinusitis"
                                                                • Go to diagnostic criteria
                                                                • Go to first line treatment: Amoxicillin (patient allergic)
                                                                • Find alternative: Macrolide-Clarythromycin
                                                                • Check for interactions
                                                                • Enter Clarithromycin/Pimozide - Note: "dangerous interaction"
                                                                • Choose another Macrolide: Azythromycin
                                                                • Recheck interaction - None noted
                                                                • Go to Azithromycin drug dosing and dosing calculator
                                                                • Enter weight in pounds
                                                                • Correct dose displayed in milligrams and milliliters
Table 4. Non-Integrated vs. Integrated PDA Software.

and calculations. This method decreases                                pital stays are prolonged, productivity is
Type 1 medical error by eliminating medi-                              impeded and healthcare costs continue
cation name confusion, ADIs, ADEs and                                  to rise due to ongoing medical error. De-
dosing miscalculations; Type 2 medical er-                             spite presidential speeches, congressional
ror by containing up-to-date peer reviewed                             hearings, the introduction of legislation
treatment plans, protocols and appropriate                             and media attention to the initial IOM re-
medication choices; and Type 3 medical                                 port, studies reveal that today patients are
error by improving efficiency and decreas-                             no safer than they were in 1999 when the
ing “fear-flawed” plans.                                               report was first published.27 It is impera-
                                                                       tive that physicians and healthcare systems
                                                                       address the medical error crisis. Given to-
Conclusion                                                             day’s advances in technology and the need
                                                                       for effective and efficient point-of-care so-
As professionals and public servants                                   lutions, PDA and integrated software solu-
scramble for solutions, lives are lost, hos-                           tions are clear choices to decrease medical

44 Harvard Health Policy Review
                                  ROSENBLOOM AND RAMSDELL: PDAS AND MEDICAL ERROR

error and cost, and improve overall patient                           errors in health care: translating research into practice.
                                                                      2000; AHRQ Publication No. 00-PO58. Available at
care.                                                        Accessed
                                                                      October 16, 2003.
                                                                16.   Morris AH. Protocol management of adult respira-
                                                                      tory distress syndrome. New Horiz. 1993;1(4):593-602.
References                                                      17.   Leape LL, Cullen DJ, Clapp MD, et al. Pharmacist par-
                                                                      ticipation on physician rounds and adverse drug events
1. Kohn L, Corrigan J, Donaldson M, eds. To err is human:             in the Intensive Care Unit. JAMA. 1999;282(3):267-
    building a safer health system. Washington DC: National           270.
    Academies Press; 1999. Institute of Medicine: Com-          18.   Kaushal R, Barker KN, Bates DW. How can informa-
    mittee on Quality of Health Care in America.                      tion technology improve patient safety and reduce
2. Liang BA. Promoting patient safety through reducing                medication errors in children’s health care. Arch Pedi-
    medical error: a paradigm of cooperation between                  atri Adoles Med. 2002;155(9):990-991.
    patient, physician and attorney. So Ill Univ Law J.         19.   Patient safety initiatives: Statement of the Joint Commission on
    2000;24(3):541-568                                                Accreditation of Healthcare Organizations before the Commit-
3. Kaushal R, Bates DW, Landrigan C, et al. Medication                tee on Health, Education, Labor and Pensions, U.S. Senate
    errors and adverse drug events in pediatric inpatients.           and the Subcommittee on Labor, Health and Human Services,
    JAMA. 2001;285(16): 2114-2120.                                    and Education of the Senate Committee on Appropriations,
4. McDonald CJ, Weiner M, Hui SL. Deaths due to medi-                 106th Cong, 2nd Sess (2000) (statement of Dennis
    cal errors are exaggerated in Institute of Medicine re-           O’Leary, president, JCAHO).
    port. JAMA. 2000;284:93.                                    20.   Institute of Medicine: Committee on Quality of
5. Thomas EJ, Studdert DM, Newhouse JP, et al. Costs                  Health Care in America. Crossing the quality chasm: a new
    of medical injuries in Utah and Colorado. Inquiry.                health system for the 21st century. Washington, DC: Na-
    1999;36:255-264.                                                  tional Academies Press; 2001.
6. Cabana MD, Rand CS, Powe NR, et al. Why don’t phy-           21.   Chin T. Doctors pull plug on paperless system. Am
    sicians follow clinical practice guidelines: a framework          Med News [online]. 2003. Available at http://www.
    for improvement. JAMA. 1999; 282(15):1458-1465.         
7. Cranney M, Warren E, Barton S, Gardner K, Walley                   Accessed October 23, 2003.
    T. Why do GPs not implement evidence-based guide-           22.   Langdorf MI, Hart M, Fox JC, et al. Man vs Machine:
    lines: a descriptive study. Fam Pract. 2001;18(4):359-            Comparison of Physician Knowledge of Drug Inter-
    363.                                                              actions with a Computerized Database in a Univer-
8. Cline DM, Welch KJ, Cline LS, Brown CK. Physician                  sity Hospital Emergency Department [abstract]. Ann
    compliance with advanced cardiac life support guide-              Emerg Med. 1998; 32(suppl. 7, pt. 2):3.
    lines. Ann Emerg Med . 1995;25(1):52-57.                    23.   Bates DW, Miller EB, Cullen DJ et al. Patient risk fac-
9. Malone DC, Shaban HM. Adherence to ATS guidelines                  tors for adverse drug events in hospitalized patients.
    for hospitalized patients with community-acquired                 Arch Intern Med. 1999;159(21):2553-2560.
    pneumonia. Ann Pharmacother. 2004;35(10):1180-1185.         24.   Bontempo. Electronic vs Printed Sources [abstract].
10. Suchyta MR, Dean NC, Narus S, Hadlock CJ. Ef-                     ACEP: Scientific Assembly. October 1998.
    fects of a practice guideline for community-acquired        25.   Siegal D. New technology and the practice of medi-
    pneumonia in an outpatient setting. Am J Med.                     cine: what are the medical-legal risks physicians face.
    2001;110(4):306-309.                                              ACEP News. 2002;21(3):4-5.
11. Young PC. The management of febrile infants by pri-         26.   Peterson LA, Brennan TA, O’Neil AC, Cook EF, Lee
    mary-care pediatricians in Utah: comparison with pub-             TH. Does housestaff discontinuity of care increase
    lished practice guidelines. Pediatrics. 1995;95 (5):623-          the risk for preventable adverse events. Ann Intern Med.
    627.                                                              1994;121(11):866-872.
12. Bazarian JJ, Veenema T, Brayer AF, Lee E. Knowledge         27.   Boodman SG. No End To Errors. Washington Post, 3
    of concussion guidelines among practitioners caring               December 2002; HE01.
    for children. Clin Pediatr. 2001; 40(4):207-212.
13. Shekelle PG, Ortiz E, Rhodes S, Morton SC, Eccles
    MP, Grimshaw JM, Woolf SH. Validity of the agency
    for healthcare research and quality clinical practice       Additional References
    guidelines: how quickly do guidelines become out-
    dated. JAMA. 2001;286(12):1461-1467.
                                                                      Adams JG, Bohan JS. System contributions to error.
14. Liang BA. Adverse event of unaddressed medical er-
                                                                      Acad Emerg Med. 2000;7(11):1189-93.
    ror: identifying and filling the holes in the health-care
    and legal systems. J Law Med Ethics. 2001;29(3/4):346-
                                                                      American Hospital Association’s Quality Advisory.
                                                                      Prescriptions for safety. 15 February 2000. Available
15. Agency for Healthcare and Quality Research. Reducing

                                                                                         Vol. 5, No. 1, Spring 2004 45

  at             physician performance and patient outcomes: a sys-
  safety20015.asp Accessed May 15,2001.                          tematic review. JAMA. 1998;280(15):1339-46.

  Bates DW, Spell N, Cullen DJ, Burdick E, Laird N, Pe-          Kaushal R, Shojania KG, Bates DW. Effects of com-
  tersen LA, Small SD, Sweitzer BJ, Leape LL. The costs          puterized physician order entry and clinical decision
  of adverse drug events in hospitalized patients. Ad-           support systems on medication safety: a systematic
  verse Drug Events Prevention Study Group. JAMA.                review. Arch Intern Med. 2003;23;163(12):1409-16.
                                                                 Leape LL. Institute of medicine medical error figures
  Bates DW, Gawande AA. Improving safety with infor-             are not exaggerated. JAMA. 2000;284(1):95-97.
  mation technology. N Engl J Med. 2003;348(25):2526-
  2534.                                                          Raschke RA, Gollihare B, Wunderlich TA, Guidry
                                                                 JR, Leibowitz AI, Peirce JC, et al. A computer alert
  Bukata WR, ed. Error in Medicine- another point of             system to prevent injury from adverse drug events:
  view. Emerg Med Acute Care Ess. 2003;27(2).                    development and evaluation in a community teaching
                                                                 hospital. JAMA. 1998;280(15):1317-20. Erratum in:
  New study estimates eight million American families            JAMA.1999;281(5):420.
  experienced a serious medical or drug error [press re-
  lease]. The Commonwealth Fund; April 15, 2002.                 Romano R. Fatal error becomes catalyst for reform.
                                                                 The Boston Globe, 15 March 1999.
  Statement on Medical Errors: Statement of the Agency for
  Healthcare Research and Quality before the Senate Appropria-   Rosenbloom M. Medical error reduction and PDAs.
  tions Subcommittee on Labor, Health and Human Services and     Intern Ped. 2003;18(2):69-77.
  Education of the Senate Committee on Appropriations, 106th
  Cong, 1nd Sess (1999) (statement of John M. Eisen-             Schenkel S. Promoting patient safety and preventing
  berg, MD, director, AHRQ).                                     medical error in emergency departments. Acad Emerg
                                                                 Med. 2000;7(11):1204-22.
  Gandhi TK, Weingart SN, Borus J, Seger AC, Peterson
  J, Burdick E, et al. Adverse drug events in ambulatory         Teich JM, Merchia PR, Schmiz JL, Kuperman GJ,
  care. N Engl J Med. 2003;348(16):1556-64.                      Spurr CD, Bates DW. Effects of computerized physi-
                                                                 cian order entry on prescribing practices. Arch Intern
  Handler JA, Gillam M, Sanders AB, Klasco R. Defin-             Med. 2000;160(18):2741-7.
  ing, identifying, and measuring error in emergency
  medicine. Acad Emerg Med. 2000;(11):1183-8.                    Weeks WB, Bagian JP. Developing a culture of safety
                                                                 in the Veterans Health Administration. Eff Clin Pract.
  Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of              2000;3(6):270-6.
  computer-based clinical decision support systems on

46 Harvard Health Policy Review