Implementation of an IT-guided checklist to improve the quality of medication history records at hospital admission
Implementation of an IT‑guided checklist to improve the quality of medication history records at hospital admission
Tanja Huber 0 1 2
Franziska Brinkmann 0 1 2
Silke Lim 0 1 2
Christoph Schröder 0 1 2
Daniel Johannes Stekhoven 0 1 2
Walter Richard Marti 0 1 2
Richard Robert Egger 0 1 2
0 Statistical Consulting , Quantik, Berikon , Switzerland
1 Department of Surgery, Cantonal Hospital Aarau , Aarau , Switzerland
2 Hospital Pharmacy, Cantonal Hospital Aarau , Aarau , Switzerland
Background Medication discrepancies often occur at transition of care such as hospital admission and discharge. Obtaining a complete and accurate medication history on admission is essential as further treatment is based on it. Objective The goal of this study was to reduce the proportion of patients with at least one medication discrepancy in the medication history at admission by implementing an IT-guided checklist. Setting Surgery ward focused on vascular and visceral surgery at a Swiss Cantonal Hospital. Method The study was divided into two phases, before and after implementation of an IT-guided checklist. For both phases a pharmacist collected and compared the medication history (defined as gold standard) with that of the admitting physician. Medication discrepancies were subdivided in omissions and commissions, incorrect medications or dose changes, and incorrect dosage forms or strength. Main outcome measure The proportion of patients with at least one medication discrepancy in the medication history before and after intervention was assessed. Results Out of 415 admissions, 228 patients that met the inclusion criteria were enrolled in the study, 113 before and 115 patients after intervention. After intervention, medication discrepancies declined from 69.9 to 29.6% (p < 0.0001) of patients, the mean medication discrepancy per patient was reduced from 2.3 to 0.6 (p < 0.0001), and the most common error, omission of a regularly used medication, was reduced from 76.4 to 44.1% (p < 0.001). Conclusion The implementation of the IT-guided checklist is associated with a significant reduction of medication discrepancies at admission and potentially improves the medication safety for the patient.
Hospital; Medication discrepancies; Medication history; Patient safety; Quality improvement; Switzerland
• Systematic acquisition of medication history at hospital
admission potentially improves medication safety.
• IT-guided checklists have a great potential for an accurate
medication history at the time of hospital admission.
Approximately one quarter of all admissions in an acute
care hospital are related to adverse drug events [
Transition of care events such as admission, transfer within the
hospital, and discharge from the hospital contribute greatly
to adverse drug events that are associated with prolonged
hospital stay, added costs, and increased mortality [
Often, adverse drug events are the result of medication
discrepancies, that are unintended documentation errors at
transition of care [
An initial source of potential medication errors lies within
the proper acquisition of the patient’s medication history
]. A systematic review of 22 studies uncovered that
between 10 and 67% of patients had at least one medication
discrepancy at hospital admission . In 6 of these studies,
up to 59% of the found discrepancies were considered to
be clinically important [
]. The most common error in the
medication history was the omission of a regularly taken
5, 6, 8, 9
The inaccurate or incomplete recording of the
medication history at admission can lead to consequences such as
interrupted or inappropriate drug therapies, duplications, or
unforeseen drug interactions with an increased risk for
drugrelated problems that may not only persist during hospital
stay but continue after discharge [
3, 6, 10, 11
Medication reconciliation interventions were previously
shown to reduce discrepancies in the medication history that
may lead to complications and subsequently prolonged
hospital stays [
Aim of the study
The aim of the study was to reduce the proportion of patients
with at least one medication discrepancy in the medication
history at hospital admission by implementing an in-house
developed electronic checklist.
The study was approved by the National Swiss Ethics
Committee on research involving humans (#2016-00939).
The prospective study was conducted in a ward with 26 beds
focusing on vascular and visceral surgery at the
Kantonsspital Aarau, a Swiss cantonal hospital. To support physicians
in obtaining accurate and complete medication histories
of patients at admission, a multidisciplinary team of
pharmacists, physicians and nurses developed a checklist in the
electronic prescribing system (EPS) to provide a framework
and guide during the acquisition of the necessary
information (Fig. 1).
Admitted patients were enrolled during the two phases of
the study. Phase 1, before the introduction of the checklist,
lasted from November 2013 to January 2014. The effect of
the checklist on medication history accuracy was examined
during phase 2, which lasted from February 2015 to May
2015. Before implementing the change to using the
checklist, all involved surgeons had to attend a mandatory seminar
on how to use the new tool.
During both phases, the medication history for each
patient was acquired by a physician and compared to that
prepared by the same pharmacist (gold-standard). All
identified medicines were documented according to the ATC code
system. Each divergence was considered a medication
discrepancy and categorized in six groups:
Omission of a medication.
Commission of a medication (additional drug not used
Medication upon admission □ No □ Yes
Drug allergies and intolerances
□ □ drug allergies
□ rash □ dyspnea □ anaphylaxis
□ others: ______________________________________________________________
□ drug intolerances
□ patient □ general practitioner □ family member □ nursing home □ homecare service □ pharmacy
□ others: ______________________________________________________________________________________________
□ incomplete: __________________________________________________________________________________________
• Incorrect medication.
• Incorrect strength.
• Incorrect dosage form.
• Incorrect dose change.
The medication management of the patients prior to
admission was categorized in five groups:
• A = at home without care service support.
• B1 = at home with nursing support.
• B2 = at home with family member support.
• B3 = at home with pharmacist support.
• I = at nursing home.
Patient admission was subdivided into three groups:
• E = elective admission.
• N = admission through the emergency department.
• SDS = admission for same day surgery.
All admitted patients (≥ 16 years), who were transferred to
the surgical ward, were included.
Patients, who were discharged from hospital within 48 h
or had been moved to another ward, before the pharmacist
could obtain the medication history, were all excluded from
the study. Likewise, patients, who were admitted for same
day surgery in phase 2 or with communication problems
and without a family member to be interviewed, were not
included in the study.
To collect the medication data of a patient, the pharmacist
used a standard structured questionnaire. When necessary,
information was assembled from the general practitioner, the
patient’s pharmacy, nursing home or a family member. To
ensure that the physicians had adequate time to collect the
medication history, the two medication lists were compared
two working days after the patient’s arrival at the ward, and
all medication discrepancies were documented.
Before starting to collect data, a power analysis was
conducted using the software R-2.15.2 (2012-10-26, Trick or
Treat). Using a Poisson regression (α = 0.05 and β = 0.05)
and assuming the intervention would reduce the
proportion of patients with at least one medication discrepancy
by 50%, a minimum enrollment of 110 patients per phase
was needed. Microsoft Excel 2010 and R-3.1.3 (2015-03-09,
Smooth Sidewalk) were used for the descriptive and
statistical analysis. Significance level for adjusted p values was set
to 0.05. In order to check whether the patient group in phase
1 differed from phase 2 univariate tests were performed.
Exact Fisher’s tests were used for nominal variables, while
Welch t-tests were performed for continuous variables.
Potential differences between medication discrepancies in
the two phases were assessed using an Exact Fisher’s test. P
values were corrected for multiple testing using the method
of Benjamini and Hochberg. Exact Fisher’s tests were
further used to calculate the odds ratios (OR) for using the
checklist and obtaining a systematic medication history.
Influence of other factors was analyzed using logistic
regression. The suitability of the model was assed using graphical
diagnostics of the fit and distributional assumptions.
228 of 415 patients admitted to the ward during the two
phases were included in this study. 113 were included in
phase 1 before intervention and 115 in phase 2 after
intervention. 187 patients were excluded for the following
reasons: 125 were discharged within less than 48 h, 39 were
admitted for same day surgery during phase 2, 15 moved to
another ward before a pharmaceutical interview could be
conducted, 7 due to communication issues, and 1 for
refusing to be interviewed.
The characteristics of the patients during phase 1 and
phase 2 are shown in Table 1. 86% were admitted during
daytime (93.8% in phase 1, 78.3% in phase 2) and 60.1%
(54.0 and 66.1%, respectively) due to emergencies (N).
While 14.2% were admitted for same day surgery (SDS)
in phase 1, no SDS admissions were included in phase 2.
The largest groups underwent vascular surgery (47.4%) or
visceral surgery (37.3%). There were slightly more men than
women in phase 1 (58%) and substantially more in phase 2
(85%). There were similar types of care services used by the
patients recruited during phase 1 and 2.
The descriptive summary of continuous variables is
summarized in Table 2. The median age of the patients in both
phases was 67 years (IQR 52-78). The median number of
regularly used medication was 5 (IQR 2-10), and the number
of discrepancies in the medication history was reduced from
a maximum of 14 down to 8 after intervention (p < 0.0001).
All other variables showed no difference.
Out of 735 medications checked before intervention,
259 (35.2%) contained discrepancies (Fig. 2). 69.9% of the
patients had at least one medication discrepancy. The most
common discrepancy in phase 1 was omission of a regularly
used medication (76.4%) followed by the commission of a
medication (6.6%). After intervention, the discrepancies
The numbers (n) and the percentage (%) of the included patients for each category are shown. Type of
admission: (E) elective, (N) emergency, (SDS) same day surgery. Type of surgeries: (g) vascular, (v)
visceral, (tr) trauma, (tx) thoracic, (a) unspecified surgeries. Type of care service: (A) at home without care
service support, (B1) at home with nursing support, (B2) at home with family member support, (B3) at
home with pharmacist support, (I) at nursing home
p = 0.54
Medication at admission
p = 0.35
Hospital stay (day)
p = 0.25
Numbers of medication
p < 0.0001
were dramatically reduced to 68 out of 677 checked
medications (10%) with a significant decrease in medication
omissions down to 44.1% (p < 0.001) and commission to
4.4% (p < 0.01). Only 29.6% of patients had at least one
For each variable the number of observations (n), the minimum (Min), the first quartile (q1), the median (x̃
), the mean (x̄), the third quartile (q3), the maximum (Max), the standard deviation (s) and the inter-quartile
range (IQR) are given. For missing data see column #NA (not available)
discrepancy after intervention (p < 0.0001), and the rate
per patient decreased from 2.3 to 0.6 (Table 2). With the
IT-guided checklist the odds ratio of having a medication
Fig. 2 Medication
discrepancies for different types of errors
before and after
intervention. Statistical significance
before and after intervention
is indicated by asterisks:
*** p < 0.001; ** p < 0.01.
There is no significant
difference in the columns without
discrepancy was 4.9 times lower (OR 4.9, 95% confidence
The most common discrepancy in both phases was
omission of a medication. In phase 1, the most discrepancies
occurred in the group of analgesics (ATC code N02) at
17.2% leading ophthalmologicals (S01) at 7.6% and
mineral supplements (A12) at 7.1%. After the implementation
of the IT-guided checklist, the medication discrepancies for
analgesics were reduced from 34 to 4, for ophthalmologicals
from 15 to 8 and for mineral supplements from 14 to 1.
Using logistic regression, various factors were examined
for their influence on the OR (Table 3). There were two
significant parameters. On one hand, the odds for a medication
discrepancy were reduced by 74% by using the checklist.
On the other hand, there was a 1.7% decrease in the odds of
having a medication discrepancy for each additional year of
age (p = 0.015).
Age was compared to the type of care service to see if
there is an association between these two factors. The
combination showed a slight trend but no statistical significance.
The acquisition of the medication history at admission was
found to be a major source for medication discrepancies
5, 6, 8, 9
]. By implementing an IT-guided checklist, our
findings show comparable results to other intervention
], indicating that medication
discrepancies can be reduced significantly by reconciliation. One
study achieved a decrease of postoperative medication
discrepancies from 40.2% of the patients in the standard
care arm to 20.3% in the intervention arm [
study reduced the number of clinically relevant medication
discrepancies per patient with a web-based intervention
from 1.44 to 1.05 [
]. In this study, the mean number
of discrepancies per patient was reduced from 2.3 to 0.6
(Table 2). A possible explanation for differences between
the studies could be found in how a medication
discrepancy is defined. The medication discrepancies in our study
were not classified based on their potential to cause harm.
We also did not restricted age or the number of
medications at admission like other studies [
5, 6, 8
it is not possible to holistically compare our results with
Our results revealed that by far the most common
discrepancy in both phases was omission of a medication. Our
finding of 76.4% omissions before intervention is higher if
compared to previous studies (46.6–62%) [
mentioned above, the difference could be found in differences
in methods, concepts and in the analysis of the outcomes.
Nevertheless, the implementation of the IT-guided checklist
reduced the incidence of omission significantly to 44.1%.
Among the omitted medication, analgesics were the most
frequently forgotten drug classes in the medication history.
This result is consistent with a study showing that analgesics
are one of three categories of drugs that cause most
]. Also often missing were ophthalmologicals and
mineral supplements. A possible explanation for mineral
supplements is provided by Cockayne et al. [
]. The study
showed that complementary and alternative medicines, so
called CAM, are rarely documented [
]. We speculate that
drug classes like ophthalmologicals and mineral
supplements are not always perceived as medication by patients.
Thus, it is necessary to specifically address the usage of
these drug classes with the patient. Since our checklist
reminded the physician to specifically address these types
of medication, it is not surprising that the implementation of
the checklist significantly reduced the medication
discrepancies of these drug classes.
The second most common discrepancy before
intervention was commission of a medication (6.6%). One
possibility that can lead to commission is copying from previous
medication lists. A previous study has found that out of 120
patients 69.2% had one medication list, 26.0% had two, 4.0%
had three and 0.8% had four lists [
]. We also
encountered patients, who had several medication lists that differed
from each other. Since the frequency of such incidents was
not documented in phase 1 or phase 2, we cannot assess,
how much they contributed to commission of a medication
and the decrease of such incidences from 6.6 to 4.4% after
The finding that a medication list does not guarantee that
a patient actually takes all the medication on the list
underlines the importance that a medication list should not just
be copied but used in combination with information from
other sources, most importantly from an interview with the
]. The concept of medication reconciliation is
that cross-referencing information from as many sources as
possible will lead to a more complete and accurate
medication history . The goal of this study was to provide the
physicians with a structured approach on how to obtain the
most accurate and complete medication history. Thus, we
can assume that the reductions in omission and
commission were largely due to the implementation of the IT-guided
The numbers for other types of discrepancies before and
after intervention such as incorrect medication, incorrect
strength, incorrect dosage form and incorrect dose change
did not change significantly (Fig. 2). The discrepancy-rate
for incorrect medication and incorrect strength was reduced,
but the numbers were too small to be significant. It is
possible that a significant effect could be seen with a larger set
The two study groups of phase 1 and 2 differed in some
aspects. For instance, the proportion of male patients and
of patients who were admitted during the night was
somewhat higher during phase 2 (Table 1). Although there was
an increase in patients for same day surgery in phase 2, they
were excluded since the attending anesthetists did not use
the checklist to record the medication history. In general,
large differences between two study groups are not
desirable, since negatively affect the comparison. In this study,
minor differences between the study groups did not appear
to affect the overall reduction of medication discrepancies
when using the checklist (p < 0.0001). In addition to the
checklist, increasing age was found to significantly reduce
the number of discrepancies. This was surprising, because
age has been shown in the literature to be a potential risk
]. However, another study has found that patients
85 years and older have a lower risk for potentially harmful
medication errors [
]. A possible explanation for age as a
protective factor could be that patients of increased age, who
are either admitted from nursing homes or have some kind
of homecare service, are admitted with a more complete and
accurate list of their medication. Our data show a slight trend
for such a connection but no statistical significance. Further
investigations are necessary to examine this hypothesis.
The study has several limitations. It was conducted with
patients of only one surgical ward and in only one hospital,
where admission process may be different from other
hospitals. A further limitation is that the obtained data from the
same day surgery patients were not excluded from phase 1.
At the end of phase 1 it was not clear that the anesthetists
would not be utilizing the checklist in phase 2. By
retrospectively excluding the same day surgery patients form phase
1, the required sample size of 110 patients would not have
As mentioned earlier, the discrepancies were not
classified based on their potential to cause harm. Further research
is needed in order to determine the impact on medication
safety by the implemented IT-guided checklist.
Our regression model was defined as the discrepancies
divided by the number of used medications. We would like
to point out that the number of discrepancies could exceed
the number of medications per patient, since a single
medication can provoke multiple discrepancies like incorrect
strength and incorrect dose change. There was one case,
where more discrepancies than regularly used medications
were made. In this case, we set the number of correct
medication to 0. It can also not be ruled out that some
medication discrepancies remained undetected, since the same
pharmacist recorded the medication history and identified
the discrepancies. However, these limitations appear to only
marginally affect the apparent result of significantly reducing
medication discrepancies by using a checklist at admission.
Discrepancies in the medication history were found to be
significantly reduced by providing a standardized, IT-guided
checklist to the physicians. Since the lack of a complete
medication history leaves patients at risk for medication
errors, the intervention applied in this study will likely
enhance the medication safety and reduce complications
due to drug-related problems.
Acknowledgements The authors thank the participating physicians
and nurses for their active effort. The authors also thank Adin
RossGillespie for his input on the logistic regression model.
1. Samoy LJ , Zed PJ , Wilbur K , Balen RM , Abu-Laban RB , Roberts M . Drug-related hospitalizations in a tertiary care medicine service of a canadian hospital: a prospective study . Pharmacother . 2006 ; 26 ( 11 ): 1578 - 86 .
2. Zed PJ , Abu-Laban RB , Balen RM , Loewen PS , Hohl CM , Brudbacher JR , et al. Incidence, severity and preventability of medication-related visits to the emergency department: a prospective study . CMAJ . 2008 ; 178 ( 12 ): 1563 - 9 .
3. Classen DC , Pestotnik SL , Evans RS , Lloyd JF , Burke JP . Adverse drug events in hospitalized patients. Excess length of stay, extra costs, and attributable mortality . JAMA . 1997 ; 277 ( 4 ): 301 - 6 .
4. Coleman EA , Smith JD , Raha D , Min SJ . Posthospital medication discrepancies: prevalence and contributing factors . Arch Intern Med . 2005 ; 165 ( 16 ): 1842 - 7 .
5. Beers MH , Munkata M , Storrie M. The accuracy of medication histories in hospital medical records of elderly persons . J Am Geriatr Soc . 1990 ; 38 ( 11 ): 1183 - 7 .
6. Lau HS , Florax C , Porsius AJ , De Boer A. The completeness of medication histories in hospital medical records of patients admitted to general internal medicine wards . Br J Clin Pharmacol . 2000 ; 49 ( 6 ): 597 - 603 .
7. Tam VC , Knowles SR , Cornish PL , Fine N , Marchesano R , Etchells EE . Frequency, type and clinical importance of medication history errors at admission to hospital: a systematic review . CMAJ . 2005 ; 173 ( 5 ): 510 - 5 .
8. Cornish PL , Knowles SR , Marchesano R , Tam V , Shadowitz S , Juurlink DN , et al. Unintended medication discrepancies at the time of hospital admission . Arch Intern Med . 2005 ; 165 ( 4 ): 424 - 9 .
9. Hellström LM , Bondesson A , Höglund P , Eriksson T. Errors in medication history at hospital admission: prevalence and predicting factors . BMC Clin Pharmacol . 2012 ; 12 : 9 .
10. Nester TM , Hale LS . Effectiveness of a pharmacist-acquired medication history in promoting patient safety . Am J Health Syst Pharm . 2002 ; 59 ( 22 ): 2221 - 5 .
11. Klopotoswska JE , Wierenga PC , Smorenburg SM , Stuijt CC , Arisz L , Kuks PF , et al. Recognition of adverse drug events in older hospitalized medical patients . Eur J Clin Pharmacol . 2013 ; 69 ( 1 ): 75 - 85 .
12. Kwan JL , Lo L , Sampson M , Shojania KG . Medication reconciliation during transitions of care as patient safety strategy: a systematic review . Ann Intern Med . 2013 ; 158 : 397 - 403 .
13. Mueller SK , Sponsler KC , Kripalani S , Schnipper JL . Hospitalbased medication reconciliation practices: a systematic review . Arch Intern Med . 2012 ; 172 ( 14 ): 1057 - 69 .
14. Schnipper JL , Liang CL , Bhan I . Effect of an electronic medication reconciliation application and process redesign on potential adverse drug events: a cluster-randomized trial . Arch Intern Med . 2009 ; 169 ( 8 ): 771 - 80 .
15. Cockayne NL , Duguid M , Shenfield GM . Health professionals rarely record history of complementary and alternative medicines . Br J Clin Pharmacol . 2005 ; 59 ( 2 ): 254 - 8 .
16. Green CF , Burgul K , Armstrong DJ . A study of the use of medicine lists in medicines reconciliation: please remember this, a list is just a list . Int J Pharm Pract . 2010 ; 18 ( 2 ): 116 - 21 .
17. Frei P , Huber LC , Simon RW , Bonani M , Lüscher TF . Insufficient medication documentation at hospital admission of cardiac patients: a challenge for medication reconciliation . J Cardiovasc Pharmacol . 2009 ; 54 ( 6 ): 497 - 501 .
18. Institute for Healthcare Improvement . Medication reconciliation review 2007 . http://www.ihi.org/explore/adesmedicationreconciliation/Pages/default.aspx. Accessed 23 Oct 2017 .
19. Gleason KM , McDaniel MR , Feinglass J , Baker DW , Lindquist L , Liss D , et al. Results of the medications at transitions and clinical handoffs (MATCH) study: an analysis of medication reconciliation errors and risk factors at hospital admission . J Gen Intern Med . 2010 ; 25 ( 5 ): 441 - 7 .
20. Climente-Marti M , Garcia-Manon ER , Artero-Mora A , JiménezTorres NV . Potential risk of medication discrepancies and reconciliation errors at admission and discharge from an inpatient medical service . Ann Phramacother . 2010 ; 44 ( 11 ): 1747 - 54 .
21. Unroe KT , Pfeiffenberger T , Riegelhaupt S , Jastrzembski J , Lokhnygina Y , Colon-Emer ic C. Inpatient medication reconciliation at admission and discharge: a retrospective cohort study of age and other risk factors for medication discrepancies . Am J Ger iatr Phar macother . 2010 ; 8 ( 2 ): 115 - 26 .
22. Pippins JR , Gandhi TK , Hamann C , Ndumele CD , Labonville SA , Diedrichsen EK , et al. Classifying and predicting errors of inpatient medication reconciliation . J Gen Intern Med . 2008 ; 23 ( 9 ): 1414 - 22 .