Series

Drug safety in oncology 1

Chemotherapy medication errors
Saul N Weingart, Lulu Zhang, Megan Sweeney, Michael Hassett

Although chemotherapy is a well established treatment modality, chemotherapy errors represent a potentially serious Lancet Oncol 2018; 19: e191–99
risk of patient harm. We reviewed published research from 1980 to 2017 to understand the extent and nature of This is the first in a Series of
medication errors in cancer chemotherapy, and to identify effective interventions to help prevent mistakes. three papers about drug safety in
oncology
Chemotherapy errors occur at a rate of about one to four per 1000 orders, affect at least 1–3% of adult and paediatric
oncology patients, and occur at all stages of the medication use process. Oral chemotherapy use is a particular area of Tufts Medical Center and Tufts
University School of Medicine,
growing risk. Our knowledge of chemotherapy errors is drawn primarily from single-institution studies at university Boston, MA, USA
hospitals and referral centres, with a particular focus on prescription orders and pharmacy practices. Although the (Prof S N Weingart MD,
heterogeneity of research methods and measures used in these studies limits our understanding of this issue, the rate L Zhang MD, M Sweeney BS);
Dana-Farber Cancer Institute,
of chemotherapy error-related injuries is generally lower than those seen in comparable studies of general medical
Boston, MA, USA
patients. Although many interventions show promise in reducing chemotherapy errors, most have little empirical (M Hassett MD); and Harvard
support. Additional research is needed to understand and to mitigate the risk of chemotherapy medication errors. Medical School, Boston, MA,
USA (M Hassett)
Introduction Given the potential risk involved, an updated review on Correspondence to:
A radical shift in the understanding of medical error medication errors that contextualises chemotherapy Prof Saul N Weingart, Tufts
Medical Center and Tufts
accompanied the publication of the 1999 Institute of specific risks and remedies could inform current University School of Medicine,
Medicine (IOM) report, To Err is Human.1 The IOM panel practice. Therefore, we decided to do a review of the Boston, MA 02111, USA
called for a transformation in the way health-care published literature to understand the extent and nature sweingart@tuftsmedicalcenter.­
org
professionals understand medical error by applying of medication errors related to chemotherapy as well as
principles from cognitive psychology and human factors, strategies for improvement. For this Series paper, we
the study of human performance in work environments. used the National Coordinating Council for Medical For more on the National
Improvements in aviation and other safety-oriented Error Reporting and Prevention definition of medication Coordinating Council see http://
www.nccmerp.org/about-
industries, such as chemical engineering, manufacturing, error as “any preventable event that may cause or lead to medication-errors
and nuclear power, showed that complex systems, rather inappropriate medication use or patient harm while the
than individual practitioners, were the primary sources medication is in the control of the health-care
of error and a target for improvement oppor­ tunities professional, patient, or consumer”.5
through simplification, standardisation, and technology.
Sentinel events in oncology, including the death of Extent and nature of chemotherapy errors
Betsy Lehman in 1994 at Boston’s Dana-Farber Cancer Early studies of chemotherapy errors relied on cancer
Institute, featured prominently in public perceptions of centre surveys. In one commonly cited study, Chen and
medical error. Previous research has shown that certain colleagues6 surveyed US bone marrow transplant centres
patients are at an increased risk of preventable harm that regarding accidental overdoses and safety practices from
has been related in part to their limited physiological 1989 to 1994, of which researchers received only
reserve, which typically include patients with acute 18 overdose reports from 15 of 115 responding centres
illnesses, comor­ bidities, multiple medications, and that had overseen an estimated 24 255 transplants, a low
malignant disease.2,3 Patients with cancer are at particular rate of 0·07%. The most common errors were attributed
risk for several reasons. First, the physiological reserves to infusion errors of cisplatin, carboplatin, busulfan,
of patients with cancer might be compromised by the cytosine arabinoside, and cyclophosphamide, or when
nature of the disease and its effects on vital organs, the cumulative drug dose was given as a daily dose.
immune function, and functional status. Second, many
antineoplastic therapies are toxic, with narrow Prescribing errors
therapeutic indices. Third, therapies might include novel Given the under-reporting of error inherent in hospital
agents or combinations, and require multiple dose surveys, chemotherapy error researchers typically review
adjustments and precise monitoring of laboratory medication orders and prescriptions to find evidence of
parameters. Finally, care is delivered over weeks or potential medication errors. Ranchon and colleagues,7
months by interprofessional care teams that might work for example, found prescription-writing errors in
in different clinical settings—a scenario that increases 540 (3·1%) of 17  150 consecutive chemotherapy pre­
the risk of miscommunication. Overall, cancer care is a scriptions reviewed in a prospective observational study
complex, tightly coupled system, in which inter­ at a French university hospital from 2006 to 2008. In the
dependencies are inherent and failures can have multivariable analysis, prescribing errors were more
catastrophic results.4 likely in patients who had had more than three injected

www.thelancet.com/oncology Vol 19 April 2018 e191

 Series

chemotherapy drugs (odds ratio [OR] 2·4, 95% CI 3959 patients—a rate of 3·1 errors per 1000 doses or
1·9–3·0), at least one dose modification (1·3, 1·0–1·7), or 3·4 errors per 100 patients.13 This rate is similar to those
protocols including carboplatin (4·5, 3·5–5·8). reported in studies8,9 examining order-writing errors. In
In another French university hospital study,8 341 (5·2%) of this study, dose errors and omitted drugs were the most
6607 antineoplastic prescriptions contained at least common mistakes made.13 Similarly, Markert and
one error. Nearly half of the errors were due to in­ colleagues14 analysed 22 216 consecutive chemotherapy
complete prescriptions, and 41% had a dose error. orders for 2337 patients at a German university hospital.
However, as in many studies that assess order writing, Chemo­ therapy errors were more common among
most errors were intercepted before administration, and inpatient than outpatient orders (4·5% vs 3·3%).
the 13 errors that did reach patients required enhanced However, of 3792 chemotherapy errors, only
monitoring in only two cases.8 three (0·08%) reached the patient.
Chemotherapy prescribing errors are common in the Ford and colleagues15 used a combined approach that
era of electronic prescribing as well, although technology included both nurse reports of medication errors and a
has introduced new types of error other than those found retrospective review of 200 randomly selected chemo­
in traditional handwritten orders. Aita and colleagues,9 in therapy orders for patients on the inpatient oncology unit
a retrospective study of chemotherapy prescribing errors of a US teaching hospital. Nurses reported 141 medication
in a computerised physician order-entry (CPOE) system administration errors (0·04% of all medication ad­
at an Italian university hospital, identified errors in ministrations, which affected 3% of admissions). Only
167 (20%) of 835 prescriptions. Most errors were due to a three errors resulted in an adverse drug event, defined in
programming error in which incorrect chemotherapy this study as a significant injury or discomfort resulting
orders were incorporated into standardised protocols. from a medication error. Errors were classified as either
Incomplete orders were also common, but only 3% of order-writing, dispensing, or administration in this paper-
errors had the potential for serious injury. Nerich and based system.15
colleagues10 reported 218 (1·5%) prescription errors National incident reporting systems draw from a
among 14  854 consecutive chemotherapy orders at a broader set of organisations and are enriched with more
French university hospital in another computerised serious events, but suffer from under-reporting. Fyhr
system from 2007 to 2008. Although two-thirds of errors and Akselsson16 examined all 60 cases reported to the
had the potential to cause harm, only eight cases (3·7%) Swedish national reporting systems from 1996 to 2008.
were potentially life-threatening. Error types included The most commonly involved drugs were fluorouracil,
dose errors, failure to acknowledge electronic alerts, and carboplatin, cytarabine, and doxorubicin. Incorrect dose
errors involving medication choice, duplicate orders, or and drug errors were the most common errors; 42%
failure to validate an order that had been entered by a (n=25) of errors occurred during prescription, and 42%
junior physician. Risk factors for severe prescription (n=25) in dispensing. Similarly, Rinke and colleagues17
errors included orders with more than two chemotherapies reviewed 310 reports from 69 facilities to the US
(OR 2·3, 95% CI 1·6–3·2), occasional users of the Pharmacopeia MEDMARX national voluntary reporting
electronic prescribing system (4·8, 2·4–9·1), and junior system from 1999 to 2004 involving chemotherapies for
physicians (1·7, 1·2–2·4). children under the age of 18 years. 85% of errors with an
identified practice setting occurred in inpatient units.
Incident and intervention reports 264 (85%) of reported errors reached the patient. Most
Because order-writing errors are often detected and errors caused no harm, but 49 (16%) incidents required
corrected before reaching the patient, researchers use monitoring or intervention. About half of errors occurred
internal voluntary safety reporting systems in hospitals to during medication administration. The most commonly
identify sources of potentially harmful errors.11 Front-line involved medications included metho­ trexate (15%),
clinicians submit reports that hospitals and clinics cytarabine (12%), and etoposide (8%), and the most
use to identify opportunities for improvement. common errors involved incorrect dose or quantity
Hospital pharmacies might also record pharmacists’ (23%), incorrect timing (23%), omissions (14%), and
interventions—ie, actions by pharmacists to clarify orders wrong administration technique or route (12%).
or correct errors. Incident and intervention reports detect Overall, studies that used incident reports and
errors that occur throughout the medication use process. pharmacy interventions to identify errors generally
For example, Serrano-Fabiá and colleagues12 used incident found error rates of three to four errors per 1000 orders
reports collected at the inpatient oncology unit of a and affected 3–17% of patients. These numbers were
Spanish teaching hospital and identified 276 medical similar to those observed in studies of ordering and
errors (21 per 1000 patient days) that had affected prescribing errors, with error rates of three to five per
225 (17%) of 1311 patients in a prospective cohort. 80% of 1000 orders in the pre-CPOE era, and 1·5 to two per
these errors were intercepted before reaching the patient. 1000 orders and 3–4% of patients with CPOE systems.
By contrast, pharmacists at a Spanish referral hospital Head-to-head comparisons of different studies are
identified 135 errors among 43 188 doses prepared for difficult given the diverse definitions and measures of

e192 www.thelancet.com/oncology Vol 19 April 2018

 Series

errors, injuries, and outcomes. However, voluntary rates tend to underestimate the risk of antineoplastic-
reporting systems are susceptible to under-reporting related adverse drug events and errors in the subset of
because staff members might be disinclined to report patients given chemotherapy.
their own mistakes, or unable to report errors that they There are few rigorous chart review studies of medical
have not witnessed. Additionally, organisations that errors in ambulatory oncology. In a prospective cohort
promote a safety culture might encourage reporting, study of outpatient chemotherapy infusion units at a
producing event rates and trends that cannot be easily US cancer centre in 2000, Gandhi and colleagues27
compared with organisations with insufficient reporting. examined 10 112 medication orders and medical records
of 1380 adults and 226 children and solicited error
Chart review studies reports. 306 (3%) medication orders had errors; 77% of
Because of these limitations, the most rigorous errors had the potential to cause harm, of which roughly
epidemiological studies of medical errors and adverse half were intercepted by pharmacists and nurses. In this
drug events rely on medical record review. Some studies study, ordering errors were more common than
use so-called triggers as indicators of potential medical dispensing or administration errors. Although the
errors. These can include out-of-bounds laboratory test chemotherapy error rate was reported as 4% in adults in
results, unexpected trans­fers to a higher level of care or this study, chemotherapy accounted for only 37% of the
administration of antidote medications, like epinephrine, total medication errors for adults. Non-chemotherapy
to facilitate nurse and physician review.18,19 medications, such as antiemetics and antihistamines,
In a classic chart review study, Brennan and colleagues2 accounted for many close call errors. The chemotherapy
abstracted 30 192 records from hospitals in New York, error rate in children was only 1% and accounted for just
NY, USA. Of the 1113 adverse events reported, including 16% of the total number of medication errors.
178 drug-related adverse events, 31 (2·8% of all events, In another outpatient study, Walsh and colleagues28
15·5% of drug-related events) were attributed to examined 1262 adult and 117 paediatric patient visits at
antitumour medications. Many of these events were three adult US oncology clinics and one paediatric
anticipated and deemed unpreventable adverse drug oncology clinic for 3 months. Among 11 908 medication
reactions, such as bone marrow suppression related to orders, the authors reported an error rate of 8·2 per
chemotherapy. Similarly, antineoplastic drugs accounted 1000 medication orders for adults (7·1% of visits) and
for 45 (1·4%) of 3325 adverse events reported in a study20 24·1 per 1000 medication orders for children (18·8% of
of 15 000 patient records from hospitals in Colorado and visits). The rates of potentially injurious errors were
Utah, but for only 0·3% of the 17 192 total adverse events. 5·0 per 1000 orders for adults and 9·9 per 1000 for
There was no evidence to suggest negligent care. children, and the rates of error-related injuries were
However, in a retrospective Australian record review 1·0 per 1000 and 4·3 per 1000 orders, respectively.
study,3 markedly elevated adverse event rates were noted, Errors were most likely to occur at the administration
in which antineoplastic drugs accounted for 22 (9·4%) stage for adults (63% of errors) and at the ordering stage
of 233 adverse drug events among 14  179 hospital for children (64%). As in the study by Gandhi and
records, of which 9% were judged to be highly colleagues,27 chemotherapy accounted for a substantial
preventable medication errors.The results of these large, percentage (40%) of medication errors. Error rates were
older studies have been replicated in Europe, Africa, an order of magnitude lower than in the study by
Asia, and Latin America, with adverse event rates of Gandhi and colleagues, especially for children. The
5–10% across inpatient populations and countries.21–25 results are difficult to compare given methodological
Recognising the contribution of medications to the differences, but differences might suggest between-site
overall rate of adverse events in hospitals, Bates and variation in event reporting or safety practices as well as
colleagues26 examined adverse drug events among differences in patients’ comorbidities. Chemotherapy
4031 adult admissions to 11 medical and surgical units error rates varied sub­stantially between these oncology
at two Boston teaching hospitals. The adverse drug clinics (from 0·3 to 5·8 per 100 visits). A common
event rate in medical units (including medical specialty source of error was confusion over two sets of active
units such as oncology) was 6·5% per patient orders—one that was written at the time of diagnosis
admission.26 Combining preventable adverse drug and one with an adjusted dose written on the day of
events and close calls (errors without injury), the overall administration. Home administration errors by family
medication error rate increased to 7·3%. Antineoplastic caregivers were common among children, at a rate of
agents accounted for 18 (7%) of 247 adverse drug events, 14·5 per 100 visits. More than half of home errors had
but only three (4%) of which were considered the potential to cause harm.
preventable, and five (3%) were considered close calls. Watts and Parsons29 prospectively studied more than
The number of events related to antineoplastic agents 20 000 chemotherapy orders from 2008 to 2011 at a single
was the fourth most frequent category for adverse drugs paediatric cancer centre. Baseline error rates of six per
events overall, and substantially lower than for 1000 visits and four per 1000 medications dispensed at
analgesics, antibiotics, and sedatives. However, these the paediatric cancer centre were within the range of

www.thelancet.com/oncology Vol 19 April 2018 e193

 Series

variation reported in Walsh and colleagues’ study, and often fatal. One review32 identified 41 cases in the
most errors were intercepted before reaching the patient. literature of vincristine that had been inadvertently
Error types included dosing or prescribing (42%), administered intrathecally. This error is often related to
treatment plan (26%), timing (12%), and pharmacy the use of medication syringes that are similar in
dispensing errors (4%). appearance and are therefore easily mistaken with one
Differences between results of studies might reflect another, with the possibility of co-administration of
varying duration of follow-up. Because cancer care is intravenous and intrathecal therapy on the same day.
provided over weeks to months, Lipitz-Snyderman and Efforts to institute safe practices, such as the use of
colleagues30 addressed a limitation in the literature in a minibags—diluted vincristine solutions administered by
retrospective cohort study of errors and adverse events infusion, rather than injection—that would eliminate
among 400 adult patients treated for breast, colorectal, or this error have been advanced, and the extent of their use
lung cancer at a US comprehensive cancer centre. Using appears to be growing.33
chart review and incident reports to identify events,
investigators followed up patients for up to a year’s Oral chemotherapy
course of therapy. 64 patients had at least one error The changing use of oral chemotherapy poses novel
resulting in harm, a rate of 0·9 per 1000 patient days. safety challenges for clinicians, patients, and family
Inpatient care represented a time of greatest risk, with a caregivers. Investigators have documented various errors
rate of harmful errors of 35·3 per 1000 patient days. The including wrong or missed doses, wrong drug, and
risk of harmful errors was higher among patients with wrong number of days supplied.34 Examples of patient
lung or colorectal cancer. The study suggests that the medication errors include accidental dosing errors of
cumulative risk of harmful errors over a course of temozolomide and continuation of treatment beyond
treatment can be substantial. the intended cycle length.35 Poor oral chemotherapy
adherence, which is considered a patient error under the
Epidemiology National Coordinating Council for Medical Error
Overall, the rate of chemotherapy errors is generally Reporting and Prevention’s definition, has been reported
lower than the rate of adverse drug events in studies of among selected populations with rates of 16–100%.36
hospitalised (5–10%) and ambulatory (25%) general High-risk groups include adoles­ cents and women on
medicine patients. This is surprising given the long-term hormonal therapies.37 Overadherence has also
vulnerability of patients with cancer, the toxicity of their been described, because of a misunderstanding of doses,
treatments, and the challenge of coordinating complex schedules, or a patient’s desire to optimise therapy,
care. In high-quality chart review studies, chemotherapy despite dose-limiting toxicities.38
error rates with the potential for harm were one to four In a 2006 survey of clinical leaders at 42 US com­
per 1000 orders and affected 1–2% of inpatients. The prehensive cancer centres, respondents from ten centres
serious injury rate due to chemotherapy errors is not reported at least one serious adverse drug event related
known. However, longitudinal studies suggest that the to oral chemotherapy in the previous year, and
cumulative error rate might be an order of magnitude respondents from 13 centres reported a close call with
higher. Medication errors affect both adults and children, the potential for serious injury.39 Few safe practices, such
although children appear to be at an increased risk. as electronic systems for prescribing oral chemotherapy,
Errors occur at all phases of medication use with different had been instituted at the time of the study, signalling
methods detecting issues at different stages, and with the need to disseminate and implement best practice
errors that have been intercepted especially common. innovations. Few centres required prescriptions to
The validity of the findings is subject to limitations include elements that are routinely used for infusion
inherent in each type of study. Even chart review studies chemotherapy such as a diagnosis, protocol number,
are limited by the quality of documentation and event cycle number, or a dose calculation. Only a third of
classification. Because studies vary in the definition of centres obtained written, informed patient consent with
errors and associated measures (eg, error per drug, per non-protocol use of oral chemotherapy, and several
dose, per patient, per visit, or per patient day), results are centres had no formal process for monitoring adherence.
impossible to reconcile. Nevertheless, the evidence Similarly, in a survey from 2012,40 only 147 (64%) of
suggests that oncology care might be safer than expected, 230 Irish community pharmacists indicated that
at least in the few teaching hospitals and referral centres prescriptions provided enough information to dispense
that have been studied to date. oral chemo­therapy safely.
A 2009 National Comprehensive Cancer Network task
Special risks force called attention to the challenges posed by oral
Intrathecal chemotherapy chemotherapies, including the need for measures to
Inadvertent intrathecal chemotherapy administration is ensure medication availability, adherence, and safe
a rare but well documented phenomenon.31 Vincristine practices.38 The task force noted challenges with
is a potent neurotoxin and intrathecal administration is supporting safe handling and home administration, and

e194 www.thelancet.com/oncology Vol 19 April 2018

 Series

potential vulnerabilities with the dispensing of oral and use of clinician double checks. Safe practices to
agents at community pharmacies. prevent verification and administration errors include
Several researchers have investigated the safety of oral measures to prevent interruptions during activities such
chemotherapy in children and adults via direct obser­ as intravenous push and pump programming.51 Proposed
vation. Taylor and colleagues41 observed how parents of safeguards for oral chemotherapy include the adoption
69 children with acute lymphoblastic leukaemia ad­ of safe prescribing standards, a meticulous approach to
ministered chemotherapy at home. Errors occurred in patient and family education about the use and safe
17 (9·9%) of 172 medication doses and could have affected handling of these drugs, and rigorous approaches to
13 (18·8%) of the 69 patients in the study. Similarly, Walsh support adherence.52
and colleagues42 observed 242 medication administrations Although centres have adopted selected best practices,
in the homes of 92 children with cancer. Nurses found there is substantial room for improvement. In a
72 medication errors in 92 administrations, 40 with the 2010 survey of National Cancer Institute-designated US
potential for injury. The most common chemotherapy cancer centres, only four of 44 centres had fully
errors were a missed or a wrong dose. Another study of implemented the 2009 ASCO and Oncology Nursing
children with acute lymphoblastic leukaemia at a tertiary Society standards.53 The greatest opportunities for im­
care centre in northern India found errors in 36 (12·5%) of provement were in the documentation of chemo­therapy
289 prescriptions due to physicians’ dose miscalculations planning, agreed-upon intervals for laboratory testing,
or administration errors at home.43 and patient education and consent before initiating oral
or infusion chemotherapy. Similarly, the ISMP published
Usual and supportive-care medications the results of self-assessments performed in 2012 by
Studies have shown higher error rates related to 352 health-care organisations in 12 countries.54 ISMP
non-chemotherapy medications than to antineoplastic identified worldwide opportunities to enhance use of
agents in patients with cancer.27–29 Little is known about WHO guidance regarding the administration of vinca
the extent to which drug interactions involving chemo­ alkaloids, management of oral chemotherapy, labelling
therapy agents and non-chemotherapy agents affect these the distal ends of chemotherapy tubing, implementation
patients,44 but the evidence for drug–drug interactions has of technology-based safeguards, and patient education.
been noted. In a written questionnaire completed by It is also important to tailor safe practice guidelines to
405 adult outpatients treated for solid tumours at a resource-poor settings that might have restricted access
Canadian cancer centre, Riechelmann and colleagues45 to expert practitioners.55
identified 240 potential drug interactions among
109 (27%) patients, including 25 potentially serious Prospective risk assessment
interactions. Potential interactions involved non- Although retrospective review and analysis of chemo­
chemotherapeutic agents, such as warfarin, antihyper­ therapy mishaps can lead to corrective actions, failure
tensive medications, corticosteroids, and anticonvulsants. mode and effects analysis, a type of prospective risk
assessment, offers a tool for understanding and
Reducing chemotherapy errors improving oral and infusion chemotherapy processes
Safe practice standards and guidelines among children and adult patients.56–60 In one academic
Drawing on best-practice recommen­dations from expert medical centre, staff used failure mode and effects
clinicians, organisations such as the Institute for Safe analysis to guide the deployment of a CPOE system
Medication Practices (ISMP and ISMP Canada), the that reduced the risk of improper chemotherapy dosing
American Society of Clinical Oncology (ASCO) with the intervals (relative risk 0·26, 95% CI 0·11–0·61),
US Oncology Nursing Society, the American Society of incorrect dosing calculations (0·09, 0·03–0·34), missing
Health-System Pharmacists, and the Clinical Oncology cumulative dose calculations (0·32, 0·14–0·77), and
Society of Australia have promulgated recommendations incomplete nursing checklists (0·51, 0·33–0·80).61
for safe practice to improve chemo­therapy safety.46–48 Best Performance improvement techniques, such as Lean and
practices include the use of unambiguous packaging and Six Sigma, can be used to identify and mitigate adverse
labelling, the use of tall man lettering (capitalising events related to chemotherapy preparation in the
certain letters within the drug name) to distinguish pharmacy.62 Multi-institutional collaborations encourage
similar looking or sounding drugs (eg, doXOrubicin vs sharing of solutions and may accelerate the pace of
daUNOrubicin), and the use of templated order sets with change.63
required fields. The move from handwritten, free-form
chemotherapy orders to pre-printed paper forms reduced Information technology
the need for clarification and duplication from 31% to Electronic order-entry systems with advanced decision
13% of orders in one US cancer centre,49 and reduced support, bar coding medication administration, and
medication errors and costs at another centre.50 Additional smart pump technology have all shown the potential to
safe practices include the use of checklists, prohibition of improve medication safety in hospitals.64 Improved
verbal orders, avoidance of ambiguous abbreviations, decision support for clinicians treating people with

www.thelancet.com/oncology Vol 19 April 2018 e195

 Series

cancer could benefit oncology practices with tools such 2013, Mattsson and colleagues71 found no significant
as drug allergy, interaction, and laboratory checking, difference in the rate of non-intercepted prescription
dose-limit warnings, tracking of cumulative lifetime errors between a unit with a CPOE system and another
doses, and calculation support for weight-based and unit with paper prescription forms (1·7% vs 1·6%;
body surface area-based dosing.65 Basic order entry p=0·49). The electronic system reduced the risk of
ensures complete prescriptions, reduces mistakes calculation errors, but introduced errors such as incorrect
related to legibility, and eliminates the need for dose reduction, full dosing despite previous dose
transcription, which, taken together, reduce error rates reduction, and failure to recalculate doses based on
by 62–79%.66,67 changes in bodyweight. The authors suggested that
However, the few studies examining the impact of effective CPOE systems require effective interfaces,
information technology on clinical outcomes in oncology robust decision support, and an ongoing need for
have shown mixed results. Small and colleagues68 found pharmacy order checking. Slow adoption of CPOE
that computerised prescribing reduced outpatient appears to be partly due to the slow development of
chemo­ therapy prescribing errors at a UK university vendor systems with the requisite features to support
hospital from 20% at baseline to 12% with computerised oncology practices.
prescriptions. Although there were reductions in dose
errors or incomplete prescriptions with CPOE, errors Patient engagement
involving the wrong stage or cycle number increased Patients and their families have been identified as
from baseline. Voeffray and colleagues69 evaluated the so-called vigilant partners in ensuring safe cancer care.72
effects of a CPOE system on chemotherapy prescribing Opportunities to encourage patients’ participation in
errors on the oncology unit of a Swiss university hospital. safety practices have been described, including a review
In this before and after study of medication errors that of medication and allergy lists,73 teamwork training,74 and
were identified prospectively by the pharmacy service, programmes to report errors and treatment-related
investigators reported 141 (15%) errors in 940 regimens at symptoms at an office kiosk tablet, or home computer.75,76
baseline and 75 (5%) errors in 1505 regimens after In a systematic review of patient-reported willingness
computerisation; however, 69 (92%) of 75 post-CPOE and intention to perform safety-related behaviours,
errors involved protocols that had not yet been Schwappach77 found that patients’ intention to participate
computerised. Most improvements involved errors with in error prevention was associated with self-efficacy,
the name and volume of the solute and drug dose. preventability of incidents, and perceived effectiveness of
Similarly, in a study of 6673 oral chemotherapy pre­ actions. These findings were similar to a subsequent
scriptions for patients treated at a US comprehensive survey78 that investigated the intention of patients who
cancer centre, most of the 395 dose-limit warnings were receiving chemotherapy to perform safety-oriented
pertained to an outdated dose ceiling for temozolomide.70 behaviours, in which patients were more likely to report
Aita and colleagues9 described programming errors a painful infusion than to ask providers to double-check
involving chemotherapy protocols in an observational the infusion or to disinfect their hands. Schwappach
study at an Italian university hospital. Similarly, in a concluded that rigorous evaluations of patient
prospective cohort study of 5767 chemotherapy orders at engagement initiatives designed to enhance patient
two Danish hospital-based cancer units from 2012 to safety are scarce.

Error type Potential intervention

Ordering Prescribing error with wrong drug, dose, time, route, or patient; drug allergy, Preprinted or templated order sets; computerised
interaction, or dose-limit error; cumulative dose exceeded; wrong or unspecified order-entry systems with advanced decision support;
number of doses or duration of therapy; errors related to weight, weight change, co-signature of trainee orders by attending physician; oral
or dosing weight; wrong cycle number or stage; failure to validate a trainee’s order chemotherapy-specific prescriptions with required or
by supervising physician; calculation error, especially involving body surface area; default fields for weight, dose, and indication
incorrect date of next cycle; unintentional drug duplication or overlap of regimens;
incorrect unit (eg, methotrexate 3 mg, rather than methotrexate 3000 mg)
Dispensing Incorrect labelling; incorrectly dispensed drug, dose, or amount Safe packaging; tall-man lettering (capitalising certain letters
within the drug name) ; pharmacy safe practices for look-
alike and sound-alike medications; home infusion pumps
with non-lethal drug supply; no interruption rules for
medication preparation or administration
Administration Unsafe handling of oncolytic medication; home misadministration; Bar code technology; smart pumps with guard rails for
over-adherence or under-adherence (oral agents) dose limits; double checks at ordering, dispensing, and
administration; patient and caregiver adherence,
education, and engagement interventions
Monitoring Failure to detect drug toxicity; failure to detect non-adherence Frequent office visits; use of symptom logs or patient portals

Table: Selected chemotherapy errors, organised by stage and potential intervention

e196 www.thelancet.com/oncology Vol 19 April 2018

 Series

Interventions An unexpected finding in the literature is the

Interventions to improve chemotherapy safety have been heightened risk in patients with cancer of errors related
proposed and trialled in various settings. Most guidelines to non-chemotherapy supportive medications and usual-
are based on expert opinion, anecdotal experience, and care medications. Non-chemotherapy errors were at least
single-institution improvement projects. We concur with as common as chemotherapy-related errors in several
Kullberg and colleagues’ conclusion79 that, despite the methodologically rigorous studies, emphasising the
absence of a robust evidence base, there is merit to the importance of medication reconciliation, robust electronic
use of incident reporting, proactive risk assessment, decision support, and close collaboration between
drug administration protocols, guidelines and checklists, oncologists and non-oncology physicians during a course
patient and provider education, and CPOE systems. of cancer treatment.
Many interventions show promise in reducing chemo­
Discussion therapy errors, although most are supported only by
More than 20 years have passed since the tragic death of expert opinion or single-institution studies. Prospective
Betsy Lehman from a chemotherapy overdose, but the risk assessment tools can help to error proof medication
study of medication errors in oncology remains in its processes and need to be used in cases with the greatest
infancy. Despite the inherent risks of chemotherapy and potential impact. Safe practice guidelines offer oppor­
the vulnerabilities of patients with cancer, the heterogeneity tunities to improve traditional infusion treatment as well
of research methods and measures limits our as home therapy and should become the standard of care.
understanding of the nature and extent of chemotherapy The adoption of safe practices is especially pressing in the
errors or the potential for interventions to mitigate risk. case of oral chemotherapy, in which many of the
We know little about chemotherapy errors in community safeguards of infusion therapy are not yet implemented.
practice or in rural or resource-poor settings. Order templates and CPOE systems improve legibility
Our knowledge of chemotherapy errors is drawn and ensure complete orders. They might introduce new
primarily from single-institution studies at university risks, but offer the opportunity to standardise care and
hospitals and referral centres, with a particular focus on incorporate advanced decision support.
prescription orders and pharmacy practice. Chemo­therapy This Series paper is limited in that our search strategy
errors occur at a rate of about one to four errors per might have inadvertently excluded meaningful con­
1000 orders, affect at least 1–3% of oncology patients tributions to the medical literature. The research literature
during a single episode of care, and occur at all stages of does not use consistent definitions of errors and injuries,
the medication use process. Certain drugs present special homogeneous study populations, or consistent methods
risks, including anthracyclines, intrathecal vinca alkaloids, to identify and ascertain errors, in turn limiting our
multidrug regimens, and oral agents administered on an ability to integrate findings across studies or generalise
intermittent basis or in extended courses of therapy. results. The heterogeneity of the literature was striking
Chemotherapy error rates vary across sites and disease given the rich research tradition of clinical oncology.
groups, a finding that might reflect varying safety practices Could the strengths of that tradition, with its emphasis on
among oncologists and nurses or reporting bias. Children
seem to be at particular risk given the prolonged, multi-
agent regimens required to treat many childhood cancers, Search strategy and selection criteria
the need for frequent dose adjustments related to changing
We searched Tufts Hirsch Health Sciences Library, Google Scholar, PubMed, MEDLINE,
weights and toxicities, and the involvement of family
Embase, ScienceDirect, Scopus, PsycINFO, and OpenGrey for publications in English from
caregivers in medication administration at home.
Jan 1, 1980, to April 30, 2017, using the terms “medication errors”, “medical errors”,
The rate of chemotherapy error-related injuries seems
“adverse events”, “preventable adverse events”, “adverse drug events”, “preventable
to be lower than in comparable studies of general
adverse drug events”, “near-misses”, or “close calls”, combined with “oncology”, “cancer
medical inpatients or in adult primary care settings,
care”, “chemotherapy”, “antineoplastic therapy”, “oral chemotherapy”, or “cancer
which might reflect the meticulous care of oncology
treatment”. Results were restricted to research articles, reviews, practice guidelines, and
providers in general and the expertise associated with a
best practices. We used filters to exclude case reports, opinion pieces, and commentaries.
narrow scope of practice. It might also be related to the
We also reviewed relevant meeting abstracts from the American Society of Clinical
multiple checks used with infusion therapy and
Oncology and the European Society for Medical Oncology. Additional articles were
collaborations among physician, pharmacists, and nurse
identified by hand searching the bibliographies of selected articles and the authors’
teams in medication delivery. Although the frequency of
personal files. References were excluded if they addressed non-human participants, or
errors appears to be lower than expected, the types of
were therapeutic clinical trials, patient education materials, or unpublished conference
potential errors are diverse (table) and the impact of a
abstracts. Because of the diversity and number of suitable references, we selected the final
chemotherapy error might be severe, a risk that is well
reference list based on each article’s originality and relevance to the scope of this Review.
known to patients with cancer.80 The increasing use of
The heterogeneous nature of the literature did not allow for the use of a standardised
oral chemotherapy raises novel concerns, because few
abstraction instrument. If multiple studies reported similar results, we selected articles
organisations have implemented the full list of safe
judged to be most relevant, recent, representative, or influential.
practice recommendations so far.

www.thelancet.com/oncology Vol 19 April 2018 e197

 Series

novel therapies, treatment-related toxicities, and cancer- 12 Serrano-Fabiá A, Albert-Marí A, Almenar-Cubells D,

Jiménez-Torres NV. Multidisciplinary system for detecting
specific outcomes obscure the specialty’s focus on use- medication errors in antineoplastic chemotherapy.
errors that result in preventable harm? Protocol violations J Oncol Pharm Pract 2010; 16: 105–12.
and deviations, the subject of audit and corrective action 13 Díaz-Carrasco MS, Pareja A, Yachachi A, Cortés F, Espuny A.
plans, represent a fundamentally different approach to Prescription errors in chemotherapy. Farm Hosp 2007; 31: 161–64.
14 Markert A, Thierry V, Kleber M, Behrens M, Engelhardt M.
patient safety than one that identifies system-related Chemotherapy safety and severe adverse events in cancer patients:
failures and fixes. Traditional markers of safe care, such strategies to efficiently avoid chemotherapy errors in in- and
as mortality and morbidity, readmissions, and adverse outpatient treatment. Int J Cancer 2009; 124: 722–28.
15 Ford CD, Killebrew J, Fugitt P, Jacobsen J, Prystas EM. Study of
drug events, are challenging to apply in settings with high medication errors on a community hospital oncology ward.
expected mortality rates, complex therapies, and expected J Oncol Pract 2006; 2: 149–54.
toxic effects. 16 Fyhr A, Akselsson R. Characteristics of medication errors with
parenteral cytotoxic drugs. Eur J Cancer Care (Engl) 2012;
21: 606–13.
Conclusion 17 Rinke ML, Shore AD, Morlock L, Hicks RW, Miller MR.
In conclusion, remarkably little is known about chemo­ Characteristics of pediatric chemotherapy medication errors in a
therapy errors outside of university hospitals and referral national error reporting database. Cancer 2007; 110: 186–95.
18 Hébert G, Netzer F, Ferrua M, Ducreux M, Lemare F, Minvielle E.
centres. Although the rate of chemotherapy error-related Evaluating iatrogenic prescribing: development of an
injuries appears to be lower than in comparable studies oncology-focused trigger tool. Eur J Cancer 2015; 51: 427–35.
of general medical patients, additional research is needed 19 Lipitz-Snyderman A, Classen D, Pfister DG, et al. Performance of a
trigger tool for identifying adverse events in oncology.
to characterise the nature and extent of harm and to J Oncol Pract 2017; 13: 223–30.
understand which interventions offer the greatest 20 Thomas EJ, Studdert DM, Burstin HR, et al. Incidence and types of
promise. Integration of safety science with clinical adverse events and negligent care in Utah and Colorado.
Med Care 2000; 38: 261–71.
practice and research represents a crucial next milestone
21 De Vries E, Ramrattan MA, Smorenburg SM, Gouma DJ,
for clinical oncology. Boermeester MA. The incidence and nature of in-hospital adverse
Contributors events: a systematic review. Qual Saf Health Care 2008; 17: 216–23.
SNW and LZ were responsible for the conception and design of the 22 Michel P, Quenon JL, Djihoud A, Tricaud-Vialle S,
study. All authors were responsible for data acquisition, analysis, or de Saraqueta AM. A French national survey of inpatients’ adverse
events prospectively assessed with ward staff.
interpretation. All authors were responsible for drafting and critical
Qual Saf Health Care 2007; 16: 369–77.
revisions, and final approval of the manuscript. All authors are
23 Aranaz-Andrés JM, Aibar-Remón C, Vitaller-Murillo J, et al.
accountable for the integrity of this work.
Incidence of adverse events related to health care in Spain: results
Declaration of interests of the Spanish National Study of Adverse Events.
We declare no competing interests. J Epidemiol Community Health 2008; 62: 1022–29.
24 Aranaz-Andrés JM, Aibar-Remón C, Limón-Ramírez R, et al.
References Prevalence of adverse events in the hospitals of five Latin American
1 Institute of Medicine. To err is human: building a safer health countries: results of the ‘Iberoamerican study of adverse events’
system. Washington: National Academies Press, 2000. (IBEAS). BMJ Qual Saf 2011; 20: 1043–51.
2 Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse events 25 Wilson RM, Michel P, Olsen S, et al. Patient safety in developing
and negligence in hospitalized patients: results of the Harvard countries: retrospective estimation of scale and nature of harm to
Medical Practice Study I. N Engl J Med 1991; 324: 370–76. patients in hospital. BMJ 2012; 344: e832.
3 Wilson RM, Runciman BW, Gibberd RW, Harrison BT, Newby L, 26 Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug
Hamilton JD. The quality in Australian health care study. events and potential adverse drug events: implications for
Med J Aust 1995; 163: 458–71. prevention. JAMA 1996; 274: 29–34.
4 Perrow C. Normal accidents: living with high-risk technologies. 27 Gandhi TK, Bartel SB, Shulman LN, et al. Medication safety in the
New York: Basic Books, 1984. ambulatory chemotherapy setting. Cancer 2005; 104: 2477–83.
5 National Coordinating Council for Medication Error Reporting and 28 Walsh KE, Dodd KS, Seetharaman K, et al. Medication errors
Prevention. About medication errors. http://www.nccmerp.org/ among adults and children with cancer in the outpatient setting.
about-medication-errors (accessed May 2, 2017). J Clin Oncol 2009; 27: 891–96.
6 Chen C, Seidel K, Armitage JO, et al. Safeguarding the 29 Watts RG, Parsons K. Chemotherapy medication errors in a
administration of high-dose chemotherapy: a national practice pediatric cancer treatment center: prospective characterization of
survey by the American Society for Blood and Marrow error types and frequency and development of a quality
Transplantation. Biol Blood and Marrow Transplant 1997; 3: 331–40. improvement initiative to lower the error rate.
7 Ranchon F, Moch C, You B, et al. Predictors of prescription errors Pediatr Blood Cancer 2013; 60: 1320–24.
involving anticancer chemotherapy agents. Eur J Cancer 2012; 30 Lipitz-Snyderman A, Pfister D, Classen D, et al. Preventable and
48: 1192–99. mitigable adverse events in cancer care: measuring risk and harm
8 Ranchon F, Salles G, Späth HM, et al. Chemotherapeutic errors in across the continuum. Cancer 2017; 123: 4728–36.
hospitalised cancer patients: attributable damage and extra costs. 31 Liu H, Tariq R Liu GL, Yan H, Kaye AD. Inadvertent intrathecal
BMC Cancer 2011; 11: 478. injections and best practice management.
9 Aita M, Belvedere O, De Carlo E, et al. Chemotherapy prescribing Acta Anaesthesiol Scand 2017; 61: 11–22.
errors: an observational study on the role of information technology 32 Hennipman B, de Vries E, Bökkerink JPM, Ball LM, Veerman AJP.
and computerized physician order entry systems. Intrathecal vincristine: 3 fatal cases and a review of the literature.
BMC Health Serv Res 2013; 13: 522. J Pediatr Hematol Oncol 2009; 31: 816–69.
10 Nerich V, Limat S, Demarchi M, et al. Computerized physician order 33 Gilbar P, Chambers CR, Larizza M. Medication safety and the
entry of injectable antineoplastic drugs: an epidemiologic study of administration of intravenous vincristine: international survey of
prescribing medication errors. Int J Med Inform 2010; 79: 699–706. oncology pharmacists. J Oncol Pharm Pract 2015; 21: 10–18.
11 France DJ, Miles P, Cartwright J, et al. A chemotherapy incident 34 Weingart SN, Toro J, Spencer J, et al. Medication errors involving
reporting and improvement system. oral chemotherapy. Cancer 2010; 116: 2455–64.
Jt Comm J Qual Patient Saf 2003; 29: 171–80.

e198 www.thelancet.com/oncology Vol 19 April 2018

 Series

35 Letarte N, Gabay MP, Bressler LR, Long KE, Stachnik JM, 58 Robinson DL, Heigham M, Clark J. Using failure mode and effects
Villano JL. Analyzing temozolomide medication errors: potentially analysis for safe administration of chemotherapy to hospitalized
fatal. J Neurooncol 2014; 120: 111–15. children with cancer. Jt Comm J Qual Patient Saf 2006; 32: 161–66.
36 Ruddy K, Mayer E, Partridge A. Patient adherence and persistence 59 Fyhr A, Ternov S, Ek Å. From a reactive to a proactive safety
with oral anticancer treatment. CA Cancer J Clin 2009; 59: 56–66. approach: analysis of medication errors in chemotherapy using
37 Partridge AH, Avorn J, Wang PS, Winer EP. Adherence to therapy general failure types. Eur J Cancer Care (Engl) 2017; 26: e12348.
with oral antineoplastic agents. J Natl Cancer Inst 2002; 94: 652–61. 60 Collins CM, Elsaid KA. Using an enhanced oral chemotherapy
38 Weingart SN, Brown E, Bach PB, et al. NCCN Task Force Report: computerized provider order entry system to reduce prescribing
Oral chemotherapy. J Natl Compr Canc Netw 2008; 6: S1–14. errors and improve safety. Int J Qual Health Care 2011; 23: 36–43.
39 Weingart SN, Flug J, Brouillard D, et al. Oral chemotherapy safety 61 Kim GR, Chen AR, Arceci RJ, et al. Error reduction in pediatric
practices at US cancer centres: questionnaire survey. BMJ 2007; chemotherapy: computerized order entry and failure modes and
334: 407. effects analysis. Arch Pediatr Adolesc Med 2006; 160: 495–98.
40 Hammond L, Marsden E, O’Hanlon N, King F, Henman MC, 62 Aboumatar HJ, Winner L, Davis R, et al. Applying Lean Sigma
Keane C. Identification of risks associated with prescribing and solutions to mistake-proof the chemotherapy preparation process.
dispensing of oral anticancer medicines in Ireland. Jt Comm J Qual Patient Saf 2010; 36: 79–86.
Int J Clin Pharm 2012; 34: 893–901. 63 Womer RB, Tracy E, Soo-Hoo W, Bickert B, DiTaranto S,
41 Taylor JA, Winter L, Geyer LJ, Hawkins DS. Oral outpatient Barnsteiner JH. Multidisciplinary systems approach to chemotherapy
chemotherapy medication errors in children with acute safety: rebuilding processes and holding the gains. J Clin Oncol 2002;
lymphoblastic leukemia. Cancer 2006; 107: 1400–06. 20: 4705–12.
42 Walsh KE, Roblin DW, Weingart SN, et al. Medication errors in the 64 Bubalo J, Warden BA, Wiegel JJ, et al. Does applying technology
home: a multisite study of children with cancer. Pediatrics 2013; throughout the medication use process improve patient safety with
131: e1405–14. antineoplastics? J Oncol Pharm Practice 2014; 206: 445–60.
43 Oberoi S, Trehan A, Marwaha RK. Medication errors on oral 65 Hammond P, Harris AL, Das SK, Wyatt JC. Safety and decision
chemotherapy in children with acute lymphoblastic leukemia in a support in oncology. Methods Inf Med 1994; 33: 371–81.
developing country. Pediatr Blood Cancer 2014; 61: 2218–22. 66 Pichon R, Zelger GL, Wacker P, Vodoz AL, Humbert J. Analysis and
44 Riechelmann RP, Saad ED. A systematic review on drug quantification of prescribing and transcription errors in a paediatric
interactions in oncology. Cancer Invest 2006; 24: 704–12. oncology service. Pharm World Sci 2002; 24: 12–15.
45 Riechelmann RP, Tannock IF, Wang L, Saad ED, Taback NA, 67 Huertas Fernández MJ, Baena-Cañada JM, Martínez Bautista MJ,
Krzyzanowska MK. Potential drug interactions and duplicate Arriola Arellano E, García Palacios MV. Impact of computerised
prescriptions among cancer patients. J Natl Cancer Inst 2007; chemotherapy prescriptions on the prevention of medication errors.
99: 592–600. Clin Transl Oncol 2006; 8: 821–25.
46 Goldspiel B, Hoffman JM, Griffith NL, et al. ASHP guidelines on 68 Small MDC, Barrett A, Price GM. The impact of computerized
preventing medication errors with chemotherapy and biotherapy. prescribing on error rate in a department of oncology/hematology.
Am J Health-Syst Pharm 2015; 72: e6–35. J Oncol Pharm Practice 2008; 14: 181–87.
47 Neuss MN, Gilmore TR, Belderson KM, et al. 2016 Updated 69 Voeffray M, Pannatier A, Stupp R, Fucina N, Leyvraz S,
American Society of Clinical Oncology/Oncology Nursing Society Wasserfallen JB. Effect of computerisation on the quality and safety
chemotherapy administration safety standards, including standards of chemotherapy prescription. Qual Saf Health Care 2006; 15: 418–21.
for pediatric oncology. Oncol Nurs Forum 2017; 44: 31–43. 70 Weingart, SN, Mattsson T, Zhu J, Shulman LN, Hassett M.
48 Carrington C, Stone L, Koczwara B, et al. The Clinical Oncological Improving electronic oral chemotherapy prescription: can we build
Society of Australia (COSA) guidelines for the safe prescribing, a safer system? J Oncol Pract 2012; 8: e168–73.
dispensing and administration of cancer chemotherapy. 71 Mattsson TO, Holm B, Michelsen H, Knudsen JL, Brixen K,
Asia Pac J Clin Oncol 2010; 6: 220–37. Herrstedt J. Non-intercepted dose errors in prescribing
49 Meisenberg BR1, Wright RR, Brady-Copertino CJ. Reduction in anti-neoplastic treatment: a prospective, comparative cohort study.
chemotherapy order errors with computerized physician order Ann Oncol 2015; 26: 981–86.
entry. J Oncol Pract 2014; 10: e5–9. 72 Schwappach DL, Wernli M. Medication errors in chemotherapy:
50 Sano HS, Waddell JA, Solimando DA Jr, Doulaveris P, Myhand R. incidence, types and involvement of patients in prevention.
Study of the effect of standardized chemotherapy order forms on A review of the literature. Eur J Cancer Care (Engl) 2010; 19: 285–92.
prescribing errors and anti-emetic cost. J Oncol Pharm Pract 2005; 73 Weingart SN, Cleary A, Seger A, et al. Medication reconciliation in
11: 21–30. ambulatory oncology. Jt Comm J Qual Patient Saf 2007; 33: 750–57.
51 Prakash V, Koczmara C, Savage P, et al. Mitigating errors caused by 74 Weingart SN, Simchowitz B, Kahlert-Eng T, et al. The “You CAN”
interruptions during medication verification and administration: campaign: teamwork training for patients and families in
interventions in a simulated ambulatory chemotherapy setting. ambulatory oncology. Jt Comm J Qual Patient Saf 2009; 35: 63–71.
BMJ Qual Saf 2014; 23: 884–92. 75 Basch E, Artz D, Dulko D, et al. Patient online self-reporting of
52 Viele CS. Managing oral chemotherapy: the healthcare toxicity symptoms during chemotherapy. J Clin Oncol 2005;
practitioner’s role. Am J Health Syst Pharm 2007; 64: S25–32. 23: 3552–61.
53 Weingart SN, Li J, Zhu J, et al. US cancer center implementation of 76 Berry DL, Blumenstein BA, Halpenny B, et al. Enhancing
ASCO/Oncology Nursing Society chemotherapy administration patient-provider communication with the electronic self-report
safety standards. J Oncol Pract 2012; 8: 7–12. assessment for cancer: a randomized trial. Clin Oncol 2011;
54 Greenall J, Shastay A, Vaida AJ, et al. Establishing an international 29: 1029–35.
baseline for medication safety in oncology: findings from the 2012 77 Schwappach DL. Engaging patients as vigilant partners in safety:
ISMP international medication safety self assessment for oncology. a systematic review. Med Care Res Rev 2010; 67: 119–48.
J Oncol Pharm Pract 2015; 21: 26–35. 78 Schwappach DL, Wernli M. Barriers and facilitators to
55 Gilbar PJ. Preventing medication errors in cancer chemotherapy chemotherapy patients’ engagement in medical error prevention.
referred to rural and remote hospitals. Aust J Rural Health 2001; Ann Oncol 2011; 22: 424–30.
9: 47–51. 79 Kullberg A, Larsen J, Sharp L. “Why is there another person’s name
56 Bonnabry P, Cingria L, Ackermann M, Sadeghipour F, Bigler L, on my infusion bag?” Patient safety in chemotherapy care—a review
Mach N. Use of a prospective risk analysis method to improve the of the literature. Eur J Oncol Nurs 2013; 17: 228–35.
safety of the cancer chemotherapy process. 80 Schwappach DL, Wernli M. Chemotherapy patients’ perceptions of
Int J Qual Health Care 2006; 18: 9–16. drug administration safety. J Clin Oncol 2010; 28: 2896–901.
57 van Tilburg CM, Leistikow IP, Rademaker CM, Bierings MB,
van Dijk AT. Health care failure mode and effect analysis: a useful
proactive risk analysis in a pediatric oncology ward.
Qual Saf Health Care 2006; 15: 58–63.

www.thelancet.com/oncology Vol 19 April 2018 e199