Expert Opinion on Drug Safety

ISSN: 1474-0338 (Print) 1744-764X (Online) Journal homepage: [Link]

Development of a list of high-risk perioperative

medications for the elderly: a Delphi method

Ke Wang, Jianghua Shen, Dechun Jiang, Xiaoxuan Xing, Siyan Zhan & Suying
Yan

To cite this article: Ke Wang, Jianghua Shen, Dechun Jiang, Xiaoxuan Xing, Siyan Zhan & Suying
Yan (2019): Development of a list of high-risk perioperative medications for the elderly: a Delphi
method, Expert Opinion on Drug Safety, DOI: 10.1080/14740338.2019.1629416

To link to this article: [Link]

View supplementary material

Accepted author version posted online: 06

Jun 2019.
Published online: 23 Jun 2019.

Submit your article to this journal

Article views: 34

View Crossmark data

Full Terms & Conditions of access and use can be found at

[Link]
EXPERT OPINION ON DRUG SAFETY
[Link]

ORIGINAL RESEARCH

Development of a list of high-risk perioperative medications for the elderly: a

Delphi method
a,b
Ke Wang , Jianghua Shena,b, Dechun Jianga,b, Xiaoxuan Xinga,b, Siyan Zhanc and Suying Yana,b
a
Department of Pharmacy, Xuanwu Hospital, The First Clinical Medical College of Capital Medical University, Beijing, China; bNational Clinical
Research Center for Geriatric Disorders, Beijing, China; cSchool of Public Health, Peking University, Beijing, China

ABSTRACT ARTICLE HISTORY

Objectives: There is a lack of direct evidence for the management of perioperative medications in elderly Received 5 March 2019
patients. Therefore, the authors aimed to develop a list of high-risk medications for the elderly population Accepted 5 June 2019
in China to provide indicators for clinicians to identify medication-related factors contributing to potential KEYWORDS
adverse events during the perioperative period. Delphi round; elderly; high-
Methods: The initial list of high-risk perioperative medications was developed by studying all the publica- risk medication; medication
tions that described specific high-risk medications and their risk profiles in the elderly. Delphi consultations list; perioperative period
were performed to form a consensus among the group of experts and the list was finalized.
Results: The expert panel consisted of 36 experts from 29 tertiary hospitals and 18 provinces or
municipalities. The consensus was reached after two Delphi rounds. Finally, a total of 86 medications of
13 medication classes and 120 screening items were included in the final list, along with perioperative risk
profiles and risk aversion recommendations for each drug.
Conclusion: This is the first study to establish a high-risk perioperative medication list in China, which can be
used as a reference for intervention and evaluation of perioperative medications for the elderly population.

1. Introduction patients, while drug management is less discussed. For exam-

ple, the American College of Surgeons (ACS) along with the
With the continuous development of medical technology, the
American Geriatrics Society (AGS) has developed a guideline in
number of surgeries performed is increasing worldwide [1].
2012, which defined nine assessment categories. For periopera-
A large proportion of elderly patients undergo surgery [2,3], and
tive medication management, they recommend Beers Criteria
are at a greater risk of postoperative morbidity and mortality
to minimize the patient risk for adverse drug reactions [8]. The
compared with patients of other age groups [4]. With multiple
most popular tools to identify potential inappropriate medica-
coexisting medical conditions, elderly people often have to take
tions (PIMs) in the elderly population are Beers Criteria and
several drugs. A study has shown that patients who take medi-
STOPP/START criteria. The criteria for PIMs for the elderly popu-
cines unrelated to surgery have a 2.7-fold increase in the risk of
lation in China was also published in 2017. In addition, perio-
postoperative complications than those who do not take such
perative medication management guidelines have been
medications [5]. In a prospective observational study, Karen, et al
published in some countries, such as the United States [9,10],
found that approximately 1 in 20 perioperative medication admin-
England [11], and Portugal [12], although these are not specific
istrations results in a medication error (ME) and/or adverse drug
for the elderly population.
event (ADE), and more than one-third of the MEs leads to ADEs [6].
To provide a reference to evaluate preoperative medica-
Both perioperative poly-medications and inappropriate withdra-
tions for elderly patients, we selected high-risk drugs that are
wal of drugs increase the risk of ADEs [7]. Above all, it is highly
administered during the perioperative period to the elderly
necessary to manage perioperative medications as well.
population through the Delphi method. We also assigned risk
Perioperative medication management for the elderly popu-
profiles and related suggestions for the drugs that required
lation mainly includes pre-operative medication reconciliation
attention. This list could be used to screen medications for
and improving post-operative medication compliance.
patients undergoing surgery during hospitalization.
However, the reference basis for medication reconciliation for
the elderly population before surgery is very obscure, and there
is a lack of direct and targeted evidence. Thus far, there has 2. Methods
been no list of high-risk drugs to be administered during the
This study was conducted from February to September 2018 in
perioperative period, especially in elderly patients. Guidelines
China. The study included two parts. First, we reviewed the
only recommend evaluating the disease status of perioperative

CONTACT Suying Yan yansuying10@[Link] Department of Pharmacy, Xuanwu Hospital, The First Clinical Medical College of Capital Medical University, No.45,
Changchun Street, Xicheng District, Beijing 100053, China; Siyan Zhan siyan-zhan@[Link] School of Public Health, Peking University, No. 38, Xueyuan
Road, Haidian District, Beijing 100191, China
Supplemental data fro this article can be accessed here.
© 2019 Informa UK Limited, trading as Taylor & Francis Group
2 K. WANG ET AL.

literature to draft an initial list of high-risk perioperative medica- provided with a questionnaire and the first round was started.
tions, which are prescribed to the elderly population. Second, we We collected some demographic details from the experts, such
used the Delphi technique to obtain consensus within a group of as the name and address of the hospitals they are associated
experts and developed the final list. The high-risk medications with, total number of professional years of experience, and their
during the perioperative period were judged based on the fol- clinical areas of practice.
lowing aspects: (i) due to deterioration of physical function and
risk of surgery, the drug can easily cause toxicity or adverse
2.3. Data collection
reactions; (ii) drugs that the elderly patients use or discontinue
during the perioperative period cause more harm than benefit, The data were collected by questionnaires, one for each round.
and (iii) drugs that can be replaced with safer alternatives with The questionnaires were pretested in a pilot study. We sent the
the development of medical technology. Ethics approval was not questionnaires to five experts including one geriatrician, one
required for creating this consensus-based list. anesthesiologist, one surgeon, and two pharmacists. We asked
them to comment on the content and layout of the question-
naires. Meanwhile, we got an idea of the time required and
2.1. Literature review fluency to fill the questionnaires, and through a pre-test, we
further improved the questionnaire. The paper questionnaires
We reviewed the literature to generate a list of possible high-risk
were distributed to the participating experts in April and
perioperative medications for the Chinese elderly population. We
August 2018. Each expert was asked to respond within 1 week
systematically searched the PubMed, EMBASE, Web of Science,
and was given reminders in case of no response.
Cochrane Library, SinoMed, China National Knowledge Infra-
During all the rounds, the experts were asked to rate their
structure (CNKI), Wanfang, VIP databases, and some guideline
opinion on a 5-point Likert scale, from 1 (strongly disagree) to 5
search websites (NGC, GIN, and NICE) to identify the literature
(strongly agree). They were also asked to self-rate themselves on
published before March 2018, regardless of the language. The
the authority (Cr) for each round，which was determined by the
following search terms were used: (‘aged’ or ‘elderly’ or ‘geriatric’
judgment criteria (Ca) for the drugs and their familiarities (Cs)
or ‘elderly’), (‘perioperative’ or ‘surgery’ or ‘surgical’ or ‘operation’ or
with the respective drugs [14]. Because there were no recognized
‘preoperative’), and (‘potentially inappropriate medication’ or
guidelines on an appropriate level of consensus [15], we used the
‘medication’). We also searched the gray literature to identify
following inclusion criteria: drugs whose mean score was ≥ 4.0, at
some related content about perioperative medications developed
least 75% of the experts rated ‘agreed’ or ‘strongly agreed’, and
by different organizations in the United States and United
had a coefficient of variance of < 0.2. If no question was raised by
Kingdom, along with a supplementary search using the Google
the experts, a consensus was reached and the drugs were
search engine.
included in the final list. If at most 25% of the experts rated
We included all publications that described specific high-risk
‘agree’ or ‘strongly agree’ and the mean score was < 3.0, the
perioperative medications and their risk profiles in the elderly
corresponding drug was excluded from further discussion.
population. We excluded publications that included the non-
elderly population, did not mention any medication or risk, and
did not describe any perioperative condition. Based on the data 2.4. Delphi rounds
obtained, we included the drugs that had perioperative risks in the
The questionnaire consisted of four parts. In the first part, we
elderly population and were available in the Chinese market.
collected the characteristics of the experts. In the second part,
we presented an initial list of high-risk perioperative medica-
tions generated from the literature review. The main perio-
2.2. Expert panel selection
perative risk profiles and our recommendations were listed.
To ensure expert opinion while drafting our final list and to make The relevant literature including guidelines, consensus, PIM
it a useful and practical one, we invited experts from various criteria, and other items required to form the recommenda-
geographical regions by conducting an internet search and used tions was provided through a summary in the questionnaire.
our team’s broad circle of acquaintances. The members of the The experts were asked to rate each item in this part. In
National Pharmaceutical Alliance also actively participated in the addition, we allowed the experts to provide their comments
process. We included geriatricians, anesthesiologists, surgeons, for improving the list. In the third part, the experts were asked
and pharmacists. All the experts in our panel were required to to add any medication that they considered to be of high-risk
fulfill the following criteria: (a) working in tertiary hospitals, (b) category perioperatively, if they were not included in the
have at least 10 years of experience in relevant fields, (c) holding initial list, along with specific considerations. In the last part
the position of deputy director or above or holding the post of of the questionnaire, the authority in the Delphi round was
deputy chairman or above in national or provincial academic self-rated by the experts.
institutions, and (d) having interest and willingness to participate The medications on which consensus was reached in the
in this study and guarantee completion of two Delphi rounds first round were not discussed further in the second round.
within the specified timeframe. After the first round, the medications regarding which no con-
We aimed to include 20 − 40 experts for stable results [13]. sensus was arrived at, and additional medications that the
The experts were contacted via e-mail, telephone, or face-to-face experts suggested and had supporting evidence, were included
meetings to explain the aim of our study, and we requested their in the second questionnaire. The comments made by the
participation. After obtaining informed consent, they were experts were considered to revise the second questionnaire.
 EXPERT OPINION ON DRUG SAFETY 3

We again asked the experts to rate each medication item in Table 2. Baseline characteristics of the experts.
the second round. The medications regarding which no con- N %
sensus was reached in the second round were excluded. Profession
Pharmacist 18 50.0
Anesthesiologist 6 16.7
2.5. Statistical analysis Geriatrician 6 16.7
Surgeon 6 16.7
Descriptive statistics were used to report the results. Calculation Age, y
40–49 15 41.7
of the mean, median, percentage of score > 4, and coefficient of 50–59 20 55.6
variation for each medication item was evaluated after each ≥60 1 2.8
round. The reliability of this study was analyzed by calculating Gender
Male 15 41.7
the expert positive coefficient, authority coefficient, and coeffi- Female 21 58.3
cient of concordance. Kendall’s W test was used to calculate the Highest level of education
coefficient of concordance of the experts, and the chi-squared Bachelor’s degree 8 22.2
Master’s degree 10 27.8
test was performed to test the significant degree of coordination Post-master’s degree (PhD) 18 50.0
among the expert opinions. The database was created with Professional title
Microsoft Excel version 16.0 (Microsoft, Seattle, WA, USA) and Director 30 83.3
Associate director 6 16.7
was double checked. All statistical analyses were performed Postgraduate experience
using SPSS version 21.0 (SPSS, Inc., Chicago, IL). 10–19 7 19.4
20–29 11 30.6
≥30 18 50.0
3. Results
3.1. Review of literature: establishment of an initial 36 questionnaires to the experts and all responses were retrieved.
selection list In the first round, the respondent rate of 89.3% (108/121) consult-
ing items was 100% (36/36), the respondent rate of 9.9% (108/121)
A total of 11,068 related articles were identified through our
items was 97.2% (35/36), and that of the other consulting items
initial search. After removing 896 duplicates and screening titles
was 86.1% (34/36). In the second round, 97.3% (36/37) consulting
and abstracts for studies which did not mention the risk of
items had a response rate of 100% (36/36), and 2.7% items (1/37)
perioperative medication in the elderly, 468 articles were
had a response rate of 97.2% (35/36). During the first round, there
included. After studying the full text of the articles, we extracted
were 16 (44.4%) experts who provided suggestions regarding the
high-risk perioperative medications and their risk profiles to form
high-risk perioperative medications for the elderly population
an initial selection list. This initial list covered 82 medications
along with their risk profiles, and 9 (25%) experts who submitted
belonging to 12 classes (Table1) with a total of 121 risks and
supplementary medications. In the second Delphi round, 11
recommendations to circumvent the risks.
(30.6%) experts provided suggestions.
The experts’ authority for self-evaluation is summarized in
3.2. Panel of experts Tables 3 and 4. In the two Delphi rounds, the average value of
the experts’ familiarity (Cs) with the drugs was > 0.70, and the
Thirty-six experts were invited to take part in the two Delphi average value of the experts’ judgment criteria (Ca) for the drugs
rounds. As presented in Table 2, our expert panel from 29 tertiary and the authority coefficient (Cr) were both > 0.80. For the 12
hospitals and 18 provinces or municipalities included six geriatri- medication classes included in the first round, there were 10
cians, six anesthesiologists, six surgeons from different specialties, classes (83.3%) with an experts’ authority > 0.70. For the 9 drug
and 18 pharmacists. Among them, the average age was 50.8 years. classes involved in the second round, there were 8 classes
The average experience in the relevant fields was 26.8 years, and (90.0%) with an expert authority coefficient of ≥ 0.7; the experts’
half of the experts had experience of > 30 years. They all were authority coefficient ranged from 0.67 − 0.87.
highly educated and had advanced professional titles.
The expert positive coefficient was represented by the respon-
dent rate of the questionnaire for each consulting items. We sent 3.3. Delphi rounds
During the first Delphi round, the experts were invited to rate their
Table 1. Initial list of medications overview. opinions on 121 items included in the initial list. Five items with
Medication classes Number content duplication and conflicts were excluded through
Nervous system medications 12 a discussion. Finally, there were 90 items that met the inclusion
Anesthetics or anesthesia-assisted medications 4
Psychiatric medications 13 criteria, and six items were revised according to the opinions of the
Cardiovascular system medications 14 experts. A consensus was arrived and 84 items were included in
Respiratory medications 1 the final list. There were 26 items that did not meet the inclusion
Digestive system medications 6
Urinary system medications 3 criteria, and the second round of selection was made after the
Hematological system medications 13 revision. A total of 16 medications submitted by the experts for
Endocrine system medications 7 supplementation and evidence-supported drugs included colchi-
Anti-infective medications 1
Antipyretic, analgesic, anti-inflammatory and anti-rheumatic 7 cine, biological agents (rituximab, etanercept, bevacizumab, rani-
medications bizumab, and Conbercept), and Chinese medicines, which were
Antiallergic medications 1 added to form 5 items in the second round.
4 K. WANG ET AL.

Table 3. The experts’ authority coefficient (Cr) in the first round. The degree of coordination between the two rounds of expert
Medication classes Cs Ca Cr opinions is expressed by the coefficient of concordance W. The
Nervous system medications 0.69 0.86 0.75 expert coefficient of concordance was 0.158 in the first round
Anesthetics or anesthesia-assisted medications 0.62 0.81 0.69 and 0.122 in the second round. The degree of coordination
Psychiatric medications 0.58 0.82 0.67
Cardiovascular system medications 0.84 0.91 0.85 among the expert opinions was significant after the χ2 test
Respiratory medications 0.80 0.88 0.81 (P < 0.05), which meant that the two rounds of expert opinions
Digestive system medications 0.79 0.89 0.81 were well coordinated and the results are reliable (Table 5)
Urinary system medications 0.69 0.85 0.75
Hematological system medications 0.66 0.85 0.73
Endocrine system medications 0.79 0.88 0.81
Anti-infective medications 0.84 0.91 0.85 3.4. Final list
Antipyretic, analgesic, anti-inflammatory and anti- 0.81 0.89 0.82
rheumatic medications After two rounds of consultation with Delphi experts, a consensus
Antiallergic medications 0.72 0.85 0.76 on 120 high-risk perioperative medications for administration in
Average value 0.74 0.87 0.77
the elderly population was reached. The final list included 86
medications belonging to 13 classes, as showed in the Sup-
Table 4. The experts’ authority coefficient (Cr) in the second round. plementary Table (available as a supplement in the online version
Medication classes Cs Ca Cr along with this article). This list focuses on the main risks of drug
Nervous system medications 0.75 0.86 0.80 administration during the perioperative period. The risk dimen-
Anesthetics or anesthesia-assisted medications 0.68 0.81 0.75 sions include delirium, cognitive impairment, fall, fluctuation in
Psychiatric medications 0.66 0.84 0.75 blood pressure, bleeding, fluctuation in blood sugar level, fluctua-
Cardiovascular system medications 0.84 0.90 0.87
Digestive system medications 0.78 0.87 0.82 tion in electrolyte levels, cardiovascular damage, and renal
Urinary system medications 0.71 0.85 0.78 damage. It can be used for screening elderly patients who will be
Hematological system medications 0.66 0.84 0.75 preparing for surgery during hospitalization to ensure medication
Endocrine system medications 0.79 0.89 0.84
Other medications 0.58 0.77 0.67 safety.
Average value 0.71 0.85 0.78

4. Discussion
Based on the results of the first-round rating, the Delphi
consultation questionnaire was revised. The questionnaire 4.1. The list
involved 32 medications belonging to 9 drug classes with Based on the available data, this study considered the poten-
a total of 37 evaluation items. After the second round, there tially inappropriate use of high-risk drugs in the elderly popu-
were 36 items that met the inclusion criteria, and a consensus lation during the perioperative period. The first Chinese drug
was reached for inclusion in the final list. There was one item risk list for the elderly population in the perioperative period
(Chinese medicine) that did not meet the inclusion criteria and was developed to fill the literature gap in this field. Combined
was excluded after discussion. The results of the Delphi rounds with the current domestic and international guidelines and
are summarized in Figure 1. consensus and after two rounds of Delphi revision of the

Figure 1. Overview of the Delphi rounds.

 EXPERT OPINION ON DRUG SAFETY 5

Table 5. The coefficient of concordance (W) of experts in each round of the Delphi covered a comprehensive range of drugs and has strong
process. clinical application and reference value.
Delphi round Items W χ2 P
Round 1 121 0.158 473.041 ＜0.001
Round 2 37 0.122 153.904 ＜0.001 4.2. Delphi process
The Delphi technique is a commonly used anonymous method
expert group, perioperative recommendations for the risk of for several rounds of letters to solicit the opinions of experts. It
each drug were ascertained to facilitate clinical practice. With gives importance to the experts and avoids the influence of
our list, clinicians will be more likely to be vigilant and identify authoritative opinions on others. The researchers summarized
patients taking high-risk medications and ensure measures to each round of results, provided feedback to the experts, and
prevent risks in elderly patients in the perioperative period. consulted the experts again. By repeated discussions, a robust
Perioperative period mostly refers to the period of time start- and reliable opinion can be formulated [15]. In this study, we
ing from entering the hospital for surgical treatment till dis- conducted two Delphi rounds to develop the final list. After
charge [16]. However, there is no clear definition of the specific drafting the initial list according to the literature research, we
length of this period. Hence, this list is suitable for drug screening consulted and discussed with relevant professional experts to
during the entire period of hospitalization. By reducing the use of further improve the initial list and optimize the expert con-
high-risk drugs and incidence of perioperative adverse events, sultation. Before the formal Delphi round, some candidate
hospitalization time and related costs can be reduced, and experts were selected for pre-investigation, and the form was
patient satisfaction can be improved. Previous studies have revised and improved based on the feedback received.
reported potentially inappropriate medications for the elderly Through the above preparatory work, smooth progress of
population focusing on high-risk medications and improvement the Delphi round was guaranteed. The electronic questionnaire
in medication safety [17] without specifying any particular time could improve the efficiency of consultation [19]. However,
period. This list focuses on the main risk profiles that could affect considering that there were many drug items that need to
perioperative outcomes. Its application range is narrower than be verified by the experts in this study, the paper consultation
that of the PIM list, and hence, the feasibility is higher. Moreover, questionnaire ensured that the experts could carefully read the
the guidelines for perioperative medication management have content, carefully consider the score, and express their opi-
pointed out the advantages and disadvantages of continuing nions. We provided references to the experts in each round of
use and stopping medications, but they have not specifically the consultation process. In the second round of consultation,
considered the risks and treatment recommendations from the in order to improve the consistency of the opinions, the
perspective of the elderly population. In this study, the contents relevant contents of the references were further provided.
of the above guidelines were combined with the expert argu- The evidence strength of the literature, however, affected the
mentation to form a drug list for the elderly population and scoring results to some extent.
a consensus was reached on the circumvention recommenda- The inclusion and exclusion criteria set by the references in
tions for clinical reference. The list focuses on providing phar- this study were a combination of the average score, the per-
macy services to elderly patients in the perioperative period to centage of a score of ≥ 4, and the coefficient of variation of the
increase help in clinical decision making. scores [14,20]. After the first round, 74.4% (90/121) of the items
After two Delphi rounds, only one drug was excluded which met the inclusion criteria, and 97.3% (36/37) of the items met
reflected the high consistency of the experts’ opinion. According the inclusion criteria after the second round. Consistent expert
to the score and the following consideration, the Chinese med- opinions were obtained. Due to the good consistency of the
icines were excluded. First, Chinese medicine is a generic term for expert opinions in the first round, the consensus drug items
a class of drugs, and assessing the perioperative risks for each were not discussed in the second round, which improved the
constituent drug of this class is difficult. Second, there is less efficiency of the Delphi round.
supporting literature, and the level of evidence is also low. The positive coefficient of the experts reflected their concern
Although Chinese medicines are not included in our list, many and attention to this study. The recovery rate of the two rounds
clinicians believe that the interactions between traditional were both 100% and the response rate of the consulting items
Chinese medicines and other drugs lead to abnormal blood were > 86.1%. This high level of enthusiasm could guarantee
coagulation and poor outcomes after surgery for which they the seriousness and reliability of the results. The expert author-
recommend the Chinese medicines to stop 1 − 2 weeks before ity coefficient was > 0.7 for 90% medication classes, which
surgery [11]. ensured the accuracy of the evaluation of the Delphi consulta-
Some drugs, such as angiotensin-converting enzyme inhi- tion. However, the expert authority for anesthetic medications,
bitors and angiotensin receptor blockers were included in psychotropic medications, and other medications (biological
the final list, although there was no agreement on whether agents and Chinese medicines) were < 0.70. The experts had
to continue these drugs during the perioperative period [18]. lesser expertise, familiarity, and clinical experience with these
By combining the 2017 EACTS, 2014 AHA/ACC, OSUWMC, three classes of drugs. The expert authority coefficient of this
ACP, UK, and Portuguese guidelines on perioperative medi- study was in the ideal range, and the reliability was high. The
cation management with the Delphi expert group’s opinion, expert opinion coordination coefficients of the two rounds
different risk aversion recommendations were considered for were significant after testing, and the results were desirable.
the elderly patients undergoing cardiac surgery and non- However, the two rounds of consultation could not reach to
cardiac surgery cardiac surgery, respectively. The final list a higher degree of coordination which might be related to
6 K. WANG ET AL.

a large number of items and the clinical experts’ differences in 4.5. Limitations
evaluation of the drugs which were beyond their professional
The Delphi method combines literature research with the
expertise [21].
opinions of expert groups, and hence, the results could be
subjective [24]. This study only dealt with the development of
4.3. Expert panel the list and no practical results are provided. The subsequent
evaluation of the application of the list is still under progress.
We invited well-known experts from 29 tertiary hospitals and 18
This list is only applicable to the management of perioperative
provinces or municipalities nationwide and integrated the the-
drugs in the elderly population in China. Due to the uneven
ory and practical experience of experts in related fields includ-
distribution of medical technology, experts from Beijing
ing geriatrics, anesthesiology, surgery, and pharmacy.
accounted for one-third of the Delphi experts. The experts
According to the number of items and the difficulty in the
from some provinces or cities were not invited, which limited
evaluation, the number of expert groups was formed. Too few
the universality of the results. After clinical applications, the
people can limit the representativeness of the subjects and the
content of this list should be updated and supplemented
reliability of the evaluation. In contrast, it is difficult to organize
regularly to ensure practicability and accuracy.
too many experts as data processing will be complicated and
the workload will increase, and hence, 15 − 50 people are
considered to be the most appropriate for a Delphi [21]. 5. Conclusion
Therefore, this study was intended to invite 20 − 40 experts,
and the number of experts who eventually participated in the In this study, we developed a high-risk medication list for the
Delphi consulting was 36. In order to ensure a high rate of elderly population in China through a combination of litera-
questionnaire recovery, the experts were made to undergo an ture research and Delphi consultation. A total of 86 medica-
intention survey. We selected the experts who were interested tions belonging to 13 classes of high-risk medications and
in this research, were willing to fill out, and can guarantee the 120 screening items for the elderly were included in the final
completion of the survey within the specified timeframe. We list. The perioperative risk profiles and risk aversion recom-
actively reminded them to fill during the Delphi consultation mendations for each drug were also listed. This list can be
process. used for interventions and assessment of the reference list.
The purpose of establishing this list is to provide objective
indicators to identify high-risk medications, intervene, and
4.4. Applications and suggestions for future research reduce the use of high-risk drugs, and thereby reduce the
The perioperative high-risk medication list for the elderly devel- incidence of perioperative adverse events in elderly patients,
oped in this study might eventually be widely used in provincial which can eventually contribute to improve surgical out-
or municipal hospitals in China. Each hospital can make appro- comes. In clinical practice, the comprehensive evaluation of
priate adjustments to the list according to their drug status. For surgical patients can be carried out by combining informa-
each elderly patient who is evaluated in the perioperative tion-based means and disease assessment tools, and the list
period, it is also necessary to individualize, fully evaluate the should be periodically updated according to the latest med-
advantages and disadvantages of high-risk drug application, ical evidence available.
and take the most appropriate clinical decisions.
The surgical outcomes and postoperative complications of
elderly patients are affected by many factors apart from drugs. Acknowledgments
Cardiological, respiratory, nutritional, and other evaluations are The authors wish to acknowledge the valuable contribution of all the
required in these elderly patients [22]. Therefore, in the clinical Delphi participants who completed the two Delphi rounds and discussion.
practice, high-risk drug screening list and other system evalua-
tion forms should be used to ensure the objectivity and effec-
tiveness of the evaluation. In addition, with the continuous Author contributions
development of information technology, electronic list, and K Wang, JH Shen and SY Yan conceived and designed the study; K Wang
hospital information system mosaic, there can be improved and XX Xing searched and screened the literature; K Wang and JH Shen
efficiency of surgical risk assessment before, during, and after drafted the initial selection list and guaranteed the smooth progress of
surgery. It can greatly improve the visibility of patient informa- Delphi round; K Wang wrote the manuscript; SY Yan and DC Jiang helped
perform the analysis with constructive discussions; K Wang and SY Zhan
tion and is beneficial for the medical staffs to take better care of
revised the manuscript. SY Yan and SY Zhan made the final approval of
the patients [23]. the version to be published, and all authors agreed to be accountable for
Future research should focus on clinical applications and eva- all aspects of the work.
luation of the list to find out whether risk screening of periopera-
tive drugs for elderly patients can help in rapid identification and
interventions, reduce the risk of medication, and improve survi- Funding
val. There are further disputes and uncertainties on whether
This study was supported by Beijing Municipal Commission of Health and
some drugs having risks need to be discontinued during the Family Planning (Nos. PXM2018_026283_000002 and PXM2017_
perioperative period. The relevant medical evidence needs to 026283_000002) and Beijing Science and Technology Commission
be continuously updated to improve the support strength. Funded Project (D181100000218002).
 EXPERT OPINION ON DRUG SAFETY 7

Declaration of interest 10. OSUWMC. Preoperative testing and medication management; 2017
[cited 2017 Dec 16]. Available from: [Link]
The authors have no relevant affiliations or financial involvement with any [Link]/pages/[Link]?k=Preoperative%20Testing%
organization or entity with a financial interest in or financial conflict with 20and%20Medication%20Management
the subject matter or materials discussed in the manuscript. This includes 11. Trust ECN. Peri-operative drug management guidelines; 2011 [cited
employment, consultancies, honoraria, stock ownership or options, expert 2017 Dec 15]. Available from: [Link]
testimony, grants or patents received or pending, or royalties. 20download%20folders/FOI/FOI%20disclosure/Clinical%20informa
tion/FOI%20Log%20920%20Dec%202011%20-%20Clinical%20Info
%20-%20Diabetes%20Guidelines%20-%20Att%[Link]
Reviewer disclosures 12. Castanheira L, Fresco P, Macedo AF. Guidelines for the manage-
Peer reviewers on this manuscript have no relevant financial or other ment of chronic medication in the perioperative period: systematic
relationships to disclose. review and formal consensus. J Clin Pharm Ther. 2011;36:446–467.
• This study formed a consensus on the management of chronic
medication in the perioperative period and can be used as
ORCID a reference.
13. Akins RB, Tolson H, Cole BR. Stability of response characteristics of
Ke Wang [Link] a Delphi panel: application of bootstrap data expansion. BMC Med
Res Methodol. 2005;5:37.
14. Chen L, Jiang L, Shen A, et al. Development of a quality instrument
References for assessing the spontaneous reports of ADR/ADE using Delphi
method in China. Eur J Clin Pharmacol. 2016;72:1135–1142.
Papers of special note have been highlighted as either of interest (•) or of 15. Keeney S, Hasson F, McKenna H. Consulting the oracle: ten lessons
considerable interest (••) to readers. from using the Delphi technique in nursing research. J Adv Nurs.
1. Weiser TG, Haynes AB, Molina G, et al. Estimate of the global 2006;53:205–212.
volume of surgery in 2012: an assessment supporting improved 16. Medical definition of perioperative. 2016 [cited 2018 Nov 5].
health outcomes. Lancet. 2015;385(Suppl 2):S11. Available from: [Link]
2. Dewan SK, Zheng SB, Xia SJ. Preoperative geriatric assessment: articlekey=22186
comprehensive, multidisciplinary and proactive. Eur J Intern Med. 17. American Geriatrics Society. Updated beers criteria for potentially
2012;23:487–494. inappropriate medication use in older adults. J Am Geriatr Soc.
3. Stewart BT, Wong E, Gupta S, et al. Surgical need in an aging 2015;2015(63):2227–2246.
population: A cluster-based household survey in Nepal. Surgery. •• The 2015 AGS Beers Criteria listed the potentially inappropri-
2015;157:857–864. ate medication in older adults and effective application of
4. Nadig SN, Wright Nunes JA. Preoperative medication management: which can lead to closer monitoring drug use.
filling in the gaps. South Med J. 2013;106:447–448. 18. Hollmann C, Fernandes NL, Biccard BM. A systematic review of
5. Kennedy JM, van Rij AM, Spears GF, et al. Polypharmacy in outcomes associated with withholding or continuing
a general surgical unit and consequences of drug withdrawal. Br angiotensin-converting enzyme inhibitors and angiotensin recep-
J Clin Pharmacol. 2000;49:353–362. tor blockers before noncardiac surgery. Anesth Analg.
6. Nanji KC, Patel A, Shaikh S, et al. Evaluation of perioperative medica- 2018;127:678–687.
tion errors and adverse drug events. Anesthesiology. 2016;124:25–34. 19. Gill FJ, Leslie GD, Grech C, et al. Using a web-based survey tool to
• This study assessed the rates of perioperative medication undertake a Delphi study: application for nurse education research.
errors and adverse drug events as percentages of medication Nurse Educ Today. 2013;33:1322–1328.
administrations, evaluated their root causes and formulated 20. Maaskant JM, Eskes A, Rijn-Bikker P, et al. High-alert medications for
targeted solutions to prevent them. pediatric patients: an international modified Delphi study. Expert
7. Bettelli G. Poly-medication and perioperative medication Opin Drug Saf. 2013;12:805–814.
management; 2017 [cited 2018 Jan 02]. Available from: https:// •• The authors conducted an international modified Delphi study
[Link]/core/books/perioperative-care-of-the-elderly to identify the high-alert medications for pediatric patients
/polymedication-and-perioperative-medication-management and validated the results with reports on medication incidents
/B23AF6973AE73F2962210ABA81ED650B in children based on the national data.
8. Chow WB, Rosenthal RA, Merkow RP, et al. Optimal preoperative 21. Guang Z. Modern epidemiological methods and applications.
assessment of the geriatric surgical patient: a best practices guide- Beijing: Beijing Medical University and Peking Union Medical
line from the American College of surgeons national surgical qual- College Press; 1994.
ity improvement program and the American geriatrics society. 22. Mohanty S, Rosenthal RA, Russell MM, et al. Optimal perioperative
J Am Coll Surg. 2012;215:453–466. management of the geriatric patient: a best practices guideline
• A best guideline from the American College of Surgeons from the American college of surgeons NSQIP and the American
National Surgical Quality Improvement Program and the geriatrics society. J Am Coll Surg. 2016;222:930–947.
American Geriatrics Society provided several aspects on the 23. Munter KH, Moller TP, Ostergaard D, et al. Implementation of an
preoperative assessment of the geriatric surgical patient. electronic checklist to improve patient handover from ward to oper-
9. Physicians A Perioperative medication management.2014[2017 ating room. J Patient Saf. 2017. DOI:10.1097/PTS.0000000000000289.
May 02; cited 2017 Dec 28]. Available from: [Link] 24. Jones J, Hunter D. Consensus methods for medical and health
org/physicians/diseases/d835/[Link] services research. BMJ (Clin Res Ed). 1995;311:376–380.