11th August 2017, Volume 130 Number 1460

Gillian Robb, Elizabeth Loe, Ashika Maharaj, Richard Hamblin, Mary E Seddon 

Adverse drug events (ADEs) and adverse drug reactions are major causes of patient morbidity and mortality, and a source of significant costs for both organisations and patients.1,2 The ability to accurately quantify medication-related harm is therefore important, but there is no gold standard for measuring such harm.

Currently, district health boards (DHBs) in New Zealand rely on voluntary reporting of incidents as the primary method to identify medication-related patient harm. Voluntary incident reporting systems have a number of limitations. Most harm goes unreported3,4 and most reporting is done by nurses rather than doctors, reflecting only part of the workforce.5 Incidents relating to the use of medication typically focus on errors, most of which do not result in harm.6 Furthermore, hospital incident reporting is unlikely to capture medication-related harm originating from the primary care or aged care sectors. The principal concern with voluntary reporting is that it cannot identify trends in ADE harms, as these may be clouded by trends in reporting rather than trends in events. It is therefore of limited value in quantifying rates of medication-related harm and driving improvement in medication safety.

In 2013, a New Zealand study used a ‘trigger tool’ (TT) to direct limited chart reviews across three DHBs.7 This study estimated a rate of 30 ADEs per 100 admissions and identified opioids and anticoagulants as the major medicine classes implicated in harm. While most events were in the lower severity harm scale, 2.5% contributed to death or permanent harm.7 This work stimulated interest in trigger tools in New Zealand as a practical approach to quantifying harm and identifying patterns of harm to inform improvement.

The trigger tool method is a systematic approach to medical record review that involves the use of pre-defined triggers. The concept was developed by the Institute for Healthcare Improvement (IHI). Initially, an ADE TT was developed in the late 1990's.8 This was subsequently modified for inclusion as part of the more comprehensive Global Trigger Tool (GTT), which has additional trigger modules (eg, surgery, intensive care).9

In 2011, the Commission initiated a trigger tool programme to promote and support the uptake of these tools in New Zealand DHBs. By 2014, 14 of 20 New Zealand DHBs had taken up the ADE TT and/or the GTT. Eight are still actively using the tool. Six DHBs discontinued using the trigger tool primarily due to resource constraints.

The purpose of this study was to identify patterns of medication-related harm from a national perspective, and to inform decisions on where to focus medication safety efforts.

Method

In this study, the same methodology was applied as that used in the first New Zealand trigger tool study, published in this journal in 2013. For a detailed explanation of the trigger tool method, see Seddon et al,7 and the Commission’s Practical Implementation Guide.10

In brief, the methodology involves a random selection of 20 medical records per month for review. Participating DHB trigger tool teams, trained and experienced in the IHI trigger tool method, review the selected charts for the ADE triggers outlined in Table 1. These triggers act as ‘flags’ for possible ADEs and if found, they prompt a standardised review of the chart to confirm if an ADE has in fact occurred. Reviewers were typically senior nurses or pharmacists and applied the same standard definitions and method as outlined by the IHI9 and the New Zealand Implementation Guide.10 The steps involved are outlined in Figure 1.

Five DHBs used the extended list of triggers in the ADE TT. One provincial DHB used only the GTT, which includes a subset of 14 of the 21 ADE triggers (see Table 1).

DHBs still active in the trigger tool programme submitted two years of anonymised data (1 July 2013 to 30 June 2015) to the Commission using a standard trigger tool template. Case weight and demographic data for the sample were obtained from each DHB’s decision support team.

Inter- and intra-rater reliability were not assessed by the individual review teams, however, the internal consistency of the data was assured to an extent, as the review teams underwent comprehensive training, used standardised definitions and followed the New Zealand Implementation Guide. There were also regular local and national review team meetings where difficult cases were discussed and definitions of harm severity agreed on.

Figure 1: Trigger tool methodology.

The standard IHI methodology includes the following:

  • DHBs randomly select a set of 20 medical records per month that meet the inclusion and exclusion criteria:
    • Inclusion:
      • Patients (age 18 years and older)
      • Length of stay at least 24 hours and formally admitted to the hospital
      • Closed and completed record (discharge summary and all coding is complete).
    • Exclusion:
      • Paediatric/neonatal patients (age <18)
      • Mental health admissions.
  • Records are reviewed by a team of two, usually nurses or pharmacists, who independently review each record and then reach agreement about identified harm and harm classification.
  • Reviewers apply a systematic approach to the record review, with a time limit of 20 minutes per record.
  • Standard triggers for either ADE TT or the medication module of the GTT (Table 1) are used when reviewing the record to flag potential harm.
  • Positive triggers initiate a more in-depth record review to determine if harm has occurred.
  • Identified harms are classified by severity (Table 2) and type (Table 3) and documented in a database to capture basic patient demographics, location of harm and length of stay data. Where possible, medications involved are recorded.
  • Harms are routinely validated by a medical reviewer.
  • Trigger tool teams typically meet monthly to discuss the classification of harm to ensure ongoing consistency and reliability of the data.

Definitions

Triggers

Table 1 outlines the triggers in both the ADE TT and the GTT.

Table 1: ADE triggers.

ADE triggers (n=5 DHBs)

GTT medication module triggers (n=1 DHB)

1

Vitamin K use

Vitamin K use

2

Antihistamine use

Antihistamine use

3

Flumazenil use

Flumazenil use

4

Naloxone use

Naloxone use

5

Anti-emetic use

Anti-emetic use

6

Over-sedation/hypotension

Over-sedation/hypotension

7

Abrupt medication stop

Abrupt medication stop

8

Raised urea/creatinine (>2 x baseline)

Raised urea/creatinine (>2 x baseline)

9

International normalised ratio (INR) >4

INR >4

10

Activated partial thromboplastin time (APTT) >100 seconds

APTT >100 seconds

11

C. difficile positive

C. difficile positive

12

Hypoglycaemia: serum glucose <3.0mmol/L

Hypoglycaemia: serum glucose <3.0mmol/L

13

Laxatives

 

14

Anti-diarrhoeal

 

15

Resonium/calcium/sodium polystyrene sulfonate

 

16

White blood count <3 x 109/L

 

17

Platelet count <50 x 109/L

 

18

Digoxin level >2nmol/L

 

19

Rash

 

20

Transferred to a higher level of care/rapid response team/arrest

Code/arrest/rapid response; transfer to higher level of care (cares module GTT)

21

Other

Other 

Harm

The definition of harm, including medication-related harm, is “unintended physical injury resulting from or contributed to, by medical care that requires additional monitoring, treatment or hospitalisation, or that results in death”.9 This definition includes all harm, whether preventable or not. The authors of the IHI tool justified this decision on the basis that preventability can change rapidly with advances in care (such as the prevention of central-line-associated bacteraemia), and if included, this would interfere with the ability to assess trends over time. Also, to ensure a more consistent measure, the focus in trigger tools is primarily on harm resulting from active care. Harm resulting from the omission of care is not specifically sought, partly because acts of omission are more subjective and less easily defined. For example, a patient with inadequately treated hypertension who subsequently suffers a stroke could be an act of omission and therefore not counted, but giving a high dose of warfarin causing bleeding would.

Definitions relevant to medication harm have been previously described.7 In brief, types of medication-related harm include unintended harm caused by the drug itself at normal doses (an adverse drug reaction) or from a medication error. While most medication errors do not result in harm,11 approximately 25% of ADEs are caused by medication errors.6

The DHB reviewers classified the type of harm in three different ways: using the NCC MERP (National Coordinating Council for Medication Error Reporting and Prevention Index) (Table 2); the Florida Harm Classification table adapted for the New Zealand setting (Table 3); and where on the patient journey that the harm occurred (Table 4).

Table 2: Harm severity categories.

Category

Description

Categories A-D

No harm

Category E

Temporary harm to the patient and required intervention

Category F

Temporary harm to the patient and required initial or prolonged hospitalisation

Category G

Permanent patient harm

Category H

Intervention required to sustain life

Category I

Patient death 

Table 3: Types of medication harm.

Events related to medication/intravenous fluids

1.     Clostridium difficile medication-associated infection

2.     IV volume overload/electrolyte imbalance

3.     Kidney damage due to contrast dye

4.     Medication-related cardiac event/arrhythmia

5.     Medication-related renal insufficiency

6.     Medication-related allergic reaction

7.     Medication-related bleeding

8.     Medication-related delirium, confusion or over-sedation

9.     Medication-related diarrhoea

10.   Medication-related glycaemic events

11.   Medication-related hypotension

12.   Medication-related nausea and vomiting

13.   Medication-related constipation

14.   Other 

Table 4: Where harm occurred.

Code

Definition

Example

Inpatient

Medication-related ADE occurred during this hospital admission.

Patient had anaphylactic reaction to chlorhexidine.

Re-admission

Medication-related ADE present on admission, related to a prior discharge, within 30 days of the index admission.

Patient admitted with constipation after being discharged on morphine.

Non-inpatient

Medication-related ADE that occurred in the community and precipitated an admission.

Patient on metoprolol admitted with a fracture due to a fall resulting from a hypotensive episode. 

Analysis

The Commission Health Quality Intelligence team undertook the analyses using aggregated data only. Confidentiality of data was assured and no individual DHBs were to be identified.

Results

Of eight DHBs who submitted data, six datasets were included: four large hospitals in main centres across both the North and South Islands, and two provincial North Island hospitals. Two of the original eight datasets did not include all the data items required for analysis and were therefore excluded. The data reported cover the two-year period 1 July 2013 to 30 June 2015.

Data from 2,659 chart reviews were included. There were 751 patients identified with medication-related harm and 1,908 with no medication-related harm. Table 5 compares the demographics between these two groups.

Table 5: Comparison across variables between those with an ADE and those with no ADE.

Variable

Patients with no ADE (N=1,908)

Patients with ADE (N=751)

Difference

P-value

Length of stay (LOS) days (mean)

4.6

8.7

4.1

0.01

Age years (mean)

49.9

57.2

7.3

<.001

Gender (F)

49%

60%

11%

<.001

Case weight

1.5

2.3

0.8

0.14

Māori

11.7%

9.9%

-1.8%

0.21

Pacific Islander (PI)

12.1%

13.1%

1.0%

0.31

Non-Māori/Non-PI

76.3%

76.9%

0.6%

0.41 

For all harm categories, those harmed were more likely to be older, female and have an increased length of stay (LOS). There were no significant differences between groups for case weight and ethnicity. A similar pattern was noted for patients with more serious harm (Categories F, G, H and I).

Across all categories (inpatient, non-inpatient and readmissions), 923 harms were identified, with 28% of patients experiencing one or more medication-related harms. Harm occurred at a rate of 34.7/100 admissions and 42.5/1,000 bed days.

Harms occurred during an inpatient stay in 65.5% of patients (n=604) and 5.5% (n=49) were associated with an inpatient stay that resulted in a readmission. Events originating in the community and precipitating a hospital admission contributed 29% (n=269).

Minor harms (Category E) accounted for 61%, however, 35% (n=325) were category F, contributing to an admission or excess LOS. Serious harm occurred in 4% (n=37), with 1.6% (n=15) contributing to permanent disability or death (Category G and I), while harm that required an intervention to sustain life accounted for 2.4% (n=22) (see Table 6).

Table 6: ADEs by harm severity and status.

Harm type

Inpatient

Non-inpatient

Readmission

All

E

484

72

4

560 (61%)

F

98

185

42

325 (35%)

G

6

7

0

13 (1.4%)

H

15

5

2

22 (2.4%)

I

1

0

1

2 (0.2%)

Total

604 (65.5%)

269 (29%)

49 (5.5%)

922 

Across all harm severity levels (excluding the ‘other’ category), the top five harms, representing 70% of the burden of medication-related harm, were: constipation, hypotension, bleeding, nausea/vomiting and delirium/confusion/over-sedation (see Figure 2).

Figure 2: Types of medication-related harm (all harm categories).

c 

The top five serious harms (categories F, G, H, I) were: bleeding, hypotension, delirium/confusion/over-sedation, constipation and IV volume overload/electrolyte imbalance. Cardiac event/arrhythmias were a close sixth cause of the more serious harms. Together these six harms represented 66% of the burden of the more serious medication-related harm.

When considering individual medicines implicated in harm (see Table 7), morphine caused 16% of harm, and other opioids (fentanyl, oxycodone, codeine and tramadol) account for a further 14%. Together these medicines account for 30% of harm and are implicated in three of the five most common harms identified in Figure 2 (constipation, nausea/vomiting and delirium/confusion/over-sedation). 

Aspirin ranked fourth in the top 10 medicines (predominantly harm from bleeding). Warfarin contributed a further 1.8%. Although not in the top 10, enoxaparin (ranked 11th) added a further 1.6%. So anticoagulation/antiplatelet agents together account for 7% of harms and are the most common cause of serious harm.

Furosemide and metoprolol, implicated in hypotension-related harm, account for 5.4% of medication-related harm. The antibiotics amoxicillin and amoxicillin/clavulanic acid accounted for 2.5% of all harm (see Table 7).

Table 7: Top 10 individual medicines implicated in harm.

1

Morphine

149

16.1%

2

Fentanyl

44

4.8%

3

Oxycodone

38

4.1%

4

Aspirin

34

3.7%

5

Furosemide

28

3.0%

6

Codeine

25

2.7%

7

Amoxicillin; amoxicillin/clavulanic acid

23

2.5%

8

Metoprolol

22

2.4%

9

Tramadol

20

2.2%

10

Warfarin

17

1.8% 

Medicines identified by the trigger tool process were grouped into categories to identify those contributing most harm overall and the more severe harms (Table 8). We devised six broad categories of medicines. Cardiovascular medicines were separated into different groupings to provide more granular information for clinicians.

Table 8: Medication implicated in patient harm by harm severity.

Medicine classes

E

F

G

H

I

Total

%

Opioids (includes tramadol)

250

33

 

9

 

292

31.64%

Anticoagulants/antiplatelet agents

25

52

 

6

1

84

9.10%

Antibiotics

40

30

 

1

 

71

7.69%

Beta-blockers, nitrates, calcium channel blockers and other antianginal agents

33

27

 

1

 

61

6.61%

Diuretics

14

23

 

 

 

37

4.01%

Other cardiovascular medicines (ACE inhibitors, ARBs, centrally acting agents, lipid lowering agents

18

15

1

2

 

36

3.9%

Not recorded/name queried

70

56

4

 

1

131

14.19%

Other (groups of medicines with less than 30 harms recorded)

112

88

8

3

 

211

22.9%

 Total

562

324

13

22

2

923

100.00% 

Discussion

This study reports on a systematic approach to measuring medication-related harm and reveals significant rates occurring both in the community and in hospitals. The study analysed results from four major metropolitan DHBs and two provincial DHBs that used the trigger tool method.

The results show that 28% of patients experienced one or more medication-related harm, the rate per 100 admissions was 34.7 and there were 42.5 adverse drug events per 1,000 bed days. ADEs were more common in older female patients, and also increased with length of hospital stay. Nearly 30% of the identified harm originated in the community, and was identified because the harm precipitated a hospital admission.

Most harm (61%) was minor in nature, however, in 37 (4%) of the cases the harm was serious, requiring life-saving interventions in 22, causing permanent harm for 13, and contributing to the death of two patients. In a further 35% the harm either precipitated or prolonged hospital stay.

The top five harms across all categories were constipation, hypotension, bleeding, nausea/vomiting and delirium/confusion/over-sedation, accounting for 70% of the burden of medication-related harm.

However, for the more serious harms, a different pattern emerged, with bleeding, hypotension, delirium/over-sedation, constipation and IV volume overload/electrolyte imbalance representing the top five.

There were six groups of medicines that caused the greatest amount of harm: opioids, anticoagulants/antiplatelet agents, antibiotics, antianginal agents (beta-blockers, nitrates, calcium channel blockers and others), diuretics and other cardiovascular medicines (ACE inhibitors, ARBs, centrally acting agents and statins).

Opioids and anticoagulants/antiplatelet agents not only accounted for 40% of all harm, they were also implicated in the most severe harm.

Compared with similar international studies,13–19 our rate of harm was in the high range, although there is wide variation in all three measures between studies: ADE as a percentage of admissions (3.4–31%), ADE/100 admissions (11.5–47.2%); and ADE/1,000 inpatient days (7–61.3%). It should be noted that study design can dramatically alter these percentages (eg, a study of elderly patients taking multiple medication is likely to dramatically increase the rates). For this reason, it is most useful that we compare our results with the previous New Zealand study,7 which had a very similar design, and reported data from 2010 to 2011.

Compared with the previous study,7 the degree and severity of ADE harm observed is consistent, although the previous study identified 15% of harm that occurred in the community, as compared with 29% in the current study.

Across both studies the classes of medicines implicated were remarkably similar. Opioids remain the medicine class responsible for most harm, accounting for 33% of all harm in the previous study compared with 32% in the current study. Anticoagulants and antiplatelet agents were also significant, accounting for 10% in the previous study and 9% in the current study. Likewise, antibiotics were similar at 9% in the previous study and 8% in the current study. In contrast, the top 10 medicines implicated in harm did vary from the previous study. While morphine was similar in both studies, warfarin dropped from 6.1% in the previous study to 1.8% in the current study.

Although unknown, we speculate that the reduction in warfarin-related harm could be associated with the increased availability and use of alternative oral anticoagulants, for which there was no trigger during the study period. Two harms associated with dabigatran were identified and in the last six months, idarucizumab has been introduced as an antidote to dabigatran-associated bleeding. The prescription of this medicine could potentially be included as an additional trigger in the future.

Fentanyl increased from 2.9% of harms in the previous study to 4.8% in the current study, and tramadol decreased from 5.2% to 2.2%. We note from the New Zealand Atlas of Healthcare Variation20 that while use of fentanyl varied 12-fold between DHBs, there was a significant overall increase in the use of fentanyl between 2011 (an average of 0.8/1,000 received fentanyl) and 2015 (average of 1.7/1,000). This may account for the increased harm seen. We cannot speculate about the reduction in harm from tramadol.

The finding of increased ADEs in females is similar to that previously reported (62% in the earlier study vs 60%). However, the mean age for women in the current study was considerably younger at 57 years compared with 65 years in the previous study.

The presence of an ADE increased the average length of stay (LOS) in both studies, however, the increase was less dramatic in the current study (4.6 days vs 7.06 days). Interestingly, the LOS of patients without an ADE was also less in the current study (8.7 days vs 10.6 days).

The major strength of our study relates to the number of patients reviewed in diverse areas of the country. The DHBs providing data included major metropolitan hospitals and smaller provincial hospitals. An additional strength was that teams involved in reviewing charts had all undergone rigorous training and had been using the method for a number of years.

Like all studies in this area, our study suffers from the subjectivity inherent in the trigger tool itself (which we attempted to mitigate with regular training and reviewing of difficult cases) and the lack of a true gold standard for medication-related harm. The change of reviewers over time can also represent a threat to reliability, although this is less likely where at least one reviewer remains consistent,21 which was the case for most of the participating DHBs.

The GTT has a more limited list of medication triggers compared with the ADE TT and may have under-reported ADEs. However, only one provincial DHB was using the GTT alone and it was unlikely to significantly impact on the rate of ADEs reported, and if anything, this would be in the direction of undercounting.

We also noted that in 14% of harms, the name of the medication was either omitted or queried. While this is a relatively small number, it highlights the importance of accurate documentation if we are to fully understand which medicines are most commonly implicated in harm.

The sheer magnitude of this unintended harm from medicines should be of interest to decision makers in two main areas. The first is economic. Although most of the harms were minor, they contribute to extended hospital stays and therefore DHB costs. The finding that a substantial number of ADEs occurred in primary care and precipitated a hospital admission is also an important economic consideration. Our data is likely to be an underestimate of the true rate of medication-related harm in primary care as we were only identifying these ADEs if they contributed to an admission. Further research is needed on ADE trigger tools in primary care to understand the factors that contribute to medication-related harm in the community, so that improvement efforts can be appropriately targeted. Preliminary work has begun on the use of trigger tools in primary care with the development of a set of triggers for the New Zealand setting22 and a trial of the trigger tool methodology as part of a primary care ‘Safety in Practice’ improvement collaborative in the Auckland region.23

The second policy area is in the type of medicines that caused harm: the high-risk medicines. If we know that these medicines cause most of the harm, then it is reasonable to concentrate our efforts nationally on understanding why, and putting resources and energy into making them safer for our patients. Following on from the results of the previous New Zealand ADE study, the Commission recently supported a national approach to decreasing harm from opioids in hospitals. This was a formative collaborative involving all DHBs and one private hospital, which ran for 18 months (November 2014 to June 2016). An independent evaluation reported positive results, with many of the participating hospitals reducing opioid-related harm between 13–74% in the targeted wards.24 This work is continuing in an effort to sustain and spread improvement.

There is potential for a similar approach to be directed at improving the management of anticoagulants and antiplatelet agents, which were identified by this and the previous ADE study, as major causes of serious patient harm.

Whether to continue with the ADE trigger tool is a question for individual DHBs. The ADE TT offers an alternative and more systematic approach to measuring medication-related harm. The first requirement for improving the ‘drug delivery process’ Bates et al wrote in 1995 is an ‘effective mechanism for systematically collecting and feeding back data about ADEs’.25 The authors also emphasised the importance of focusing on system changes to address preventable ADEs.25 This advice still applies more than 20 years on.

Concerns have been raised about the resource implications of using trigger tools, yet no DHB has questioned the cost of the voluntary reporting system, a system that misses most if not all medication-related harm7 and that has a poor record in improving care. DHBs could consider alternative approaches to use trigger tools more efficiently. A targeted approach has been used successfully to identify medication-related harm in a surgical service26 and in a selected cohort of patients over 65 years of age on five or more medicines.27 This approach provides a snapshot at a specific point in time that can engage staff and inform local improvement work.

How best to measure medication safety remains a challenge.28 Perhaps the most important message comes from a recent systematic review:29 counting numbers of ADEs is perhaps less important than understanding and characterising the types of medicines that cause harm, and using this to prioritise quality improvement activity.

Considerable work is currently underway in New Zealand to mitigate harm from medication, ranging from high-tech electronic prescribing, administration and medication reconciliation systems, to increasing partnerships of clinical pharmacists with clinical teams, and the specific programmes around high-risk medication.

Conclusion

This paper reports on medication-related harm from six DHBs across New Zealand. It confirms earlier work that such harms are common, occur both in hospitals and in the community, and are a burden for patients and our healthcare system. Considerable work is underway at local and national levels to decrease this harm, with a focus on the high-risk medicines most commonly implicated. Part of such an approach should include an effective mechanism for measuring and monitoring medication-related harm, so that changes can be assessed for their ability to lead to improvement. 

Summary

The purpose of this study was to identify the extent of medication-related harm in New Zealand and to use this information to inform decision on where to focus medication safety efforts. The study used a ‘trigger tool’, which is a method designed to identify harm from reviewing medical records, rather than relying on voluntary reporting. The results show that medication-related harms are common—for every 100 hospital admissions, over 30 patients suffer medication-related harm. While most were minor (61%), more than a third (35%) prompting admission to hospital or prolonging the hospital stay, and 1.6% were associated with permanent harm or death. Opioids and blood thinners were the medication classes responsible for most harm (40%) and the most severe harm. This information helps target medication safety improvement efforts to reduce harm.

Abstract

Aim

The purpose of this study is to identify patterns of medication-related harm from a national perspective, and to use this information to inform decisions on where to focus medication safety efforts. This study updates a 2013 study using the same methodology.

Method

District health boards (DHBs) still actively using either the Adverse Drug Event (ADE) Trigger Tool (TT) or the Global Trigger Tool (GTT), submitted two years of anonymised ADE data (1 July 2013–30 June 2015) to the Health Quality & Safety Commission (the Commission) using a standard template. Analyses were conducted using aggregated data only.

Results

Of eight DHBs who submitted data, six datasets were included, representing a total of 2,659 chart reviews. From these reviews, 923 harms were identified in 751 patients, with 28% of patients experiencing one or more harms. Harms occurred at a rate of 34.7 per 100 admissions, 42.5 per 1,000 bed days and 28% of patients experienced one or more medication-related harms. Those harmed were more likely to be older, female and have an increased length of stay. Most harms (65%) occurred during an inpatient stay, however, a substantial number (29%) originated in the community and precipitated an admission. Across all levels of severity, the most common types of medication harm were constipation, hypotension and bleeding. In the more serious harm categories, bleeding, hypotension and delirium/confusion/over-sedation were most common. Six groups of medicines caused the greatest amount of harm: opioids (including tramadol), anticoagulants/antiplatelet agents, antibiotics, antianginals (beta-blockers, nitrates, calcium channel blockers and others), diuretics and other cardiovascular medicines (angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor antagonists (ARBs), centrally acting agents and statins). Opioids and anticoagulants/antiplatelet agents not only accounted for 40% of all harm, they were implicated in the most severe harm.

Conclusion

This paper confirms earlier work that medication-related harms are common, occur both in hospitals and in the community, and are a substantial burden for patients and our healthcare system. Work is underway at local and national levels to decrease this harm, with a focus on the high-risk medicines most commonly implicated.

Author Information

Gillian Robb, Senior Advisor, Health Quality & Safety Commission, Wellington, Professional Teaching Fellow, University of Auckland; Elizabeth Loe, Medication Safety Specialist, Health Quality & Safety Commission, Wellington; Ashika Maharaj, Teaching Associate, Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne, Victoria, Australia; Richard Hamblin, Director, Health Quality Intelligence, Health Quality & Safety Commission, Wellington; Mary E Seddon, Executive Director of Medical Services, West Moreton Hospital and Health Services, Queensland, Australia.

Correspondence

Dr Mary E Seddon, Seddon Healthcare Quality Ltd.

Correspondence Email

mzseddon@gmail.com

Competing Interests

Nil.

References

  1. Stausberg J. International prevalence of adverse drug events in hospitals: an analysis of routine data from England, Germany and the USA. BMC Health Services Research. 2014; 14:125.
  2. Bouvy JC, De Bruin ML, Koopmanschap MA. Epidemiology of adverse drug reactions in Europe: a review of recent observational studies Drug Safety. 2015; 38:437–453.
  3. Westbrook JI, Li L, Lehnbom EC, et al. What are incident reports telling us? A comparative study at two Australian hospitals of medication errors identified at audit, detected by staff and reported to an incident system. International Journal for Quality in Health Care. 2015; 27(1):1–9.
  4. Classen DC, Resar RK, Griffin FA, et al. Global Trigger Tool’ Shows That Adverse Events In Hospitals May Be Ten Times Greater Than Previously Measured. Health Affairs 2011; 30(4):581–589.
  5. Evans SM, Berry JG, Smith BJ, et al. Attitudes and barriers to incident reporting: a collaborative hospital study Quality & Safety in Health Care. 2006; 15:39–43.
  6. Bates DW, Boyle DL, Vliet MBV, et al. Relationship between medication errors and adverse drug events. Journal of General Internal Medicine. 1995.
  7. Seddon ME, Jackson A, Cameron C, et al. The Adverse Drug Event Collaboration: a joint venture to measure medication-related patient harm. New Zealand Medical Journal. 2013; 126(1368):9–20.
  8. Rozich JD, Haraden CR, Resar RK. Adverse drug event trigger tool: a practical methodology for measuring medication related harm. Quality & Safety in Health Care. 2003; 12:194–200.
  9. Griffin FA, Resar RK. IHI Global Trigger Tool for Measuring Adverse Events. Cambridge, MA: Institute for Healthcare Improvement; 2009.
  10. Health Quality & Safety Commission. The Global Trigger Tool: A Practical Implementation Guide for New Zealand District Health Boards Wellington 2012.
  11. Bates DW. Medication errors: How common are they and what can be done to prevent them? Drug Safety. 1996; 15(5):303–310.
  12. Adler L, Florida Hospital. Florida Hospital Global Trigger Tool Implementation Toolkit http://www.safetyleaders.org/pages/idPage.jsp?ID=4879. Accessed March 20 2017.
  13. Cohen MM, Kimmel NL, Benage MK, et al. Medication safety programme reduces adverse drug events in a community hospital. Quality & Safety in Health Care. 2005;14:169–174.
  14. Carnevali L, Krug B, Amant F, et al. Performance of the adverse drug event tool and the global trigger tool for identifying adverse drug events: Experience in a Belgian hospital. Annals of Pharmacotherapy. 2013; 47:1414–1419.
  15. Hakkarainen KM, Hedna K, Petzfold M, Hagg S. Percentage of patients with preventable adverse drug reactions and preventability of adverse drug reactions – A Meta-Analysis. PLoS ONE. 2012; 7(3):e33236
  16. Giordani F, Rozenfeld S, Martins M. Adverse drug events identified by triggers at a teaching hospital in Brazil. Pharmacology and Toxicology. 2014; 15:71.
  17. Paiboonvong T, Tragulpiankit P, Viriyakul N, Iamchoo W, Promtasarn K, Sangviroon A. Adverse drug events identified by Thai ADEs Triger Tool in hospitalised patients in Thailand (Conference Abstract). International Journal Clinical Pharmacy. 2015; 37:179–287.
  18. Anburaj E, Villakkathala R, Mallaysamy S, Stanley W. Effectiveness of Trigger Tools to detect adverse druge events from discharge summaries: A retrospective study (Conference Abstract) Value in Health. 2016; 19(7):A823.
  19. Franklin BD, Birch S, Schachter M, Barber N. Testing a trigger tool as a method of detecting patient harm from medication errors in a UK hospital: A pilot study International Journal of Pharmacy Practice. 2010; 18:305–311.
  20. Health Quality & Safety Commission. Atlas of Healthcare Variation http://www.hqsc.govt.nz/our-programmes/health-quality-evaluation/projects/atlas-of-healthcare-variation/. Accessed March 27 2017.
  21. Mevik K, Griffin FA, Hansen TE, Deilkas E, Vonen B. Is inter-rater reliability of Global Trigger Tool results altered when members of the review team are replaced? International Journal for Quality in Health Care. 2016; 28(4):492–496.
  22. Eggleton KS, Dovey SM. Using triggers in primary care patient records to flag increased adverse event risk and measure patient safety at clinic level. New Zealand Medical Journal. 2014; 127(1390).
  23. Ko Awatea. Safety in Practice. http://koawatea.co.nz/project/safety-in-practice/. Accessed March 27, 2017.
  24. Appleton-Dyer S, Edirisuriya N, Soupen A, Boswell A. Safe Use of Opioids National Formative Collaborative Evaluation: A report for the Health Quality & Safety Commission. Auckland Synergia Ltd;2016.
  25. Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug events and potential adverse drug events: Implications for prevention. JAMA. 1995; 274(1):29–34.
  26. de Boer M, Kiewiet JS, Boeker EB, et al. A targeted method for standardised assesment of adverse drug events in surgical patients Journal of Evaluation in Clinical Practice. 2013; 19:1073–1082.
  27. Klopotowska JE, Wierenga P, Stuijt CM, et al. Adverse drug evetns in older hospitalised patients: results and reliability of a comprehensive and structured identification strategy. PLOS One. 2013; 8(8):e71045.
  28. Ng J, Scahill S, Harrison J. Getting the foundations right for the measurement of medication safety: the need for a meaningful conceptual frame. New Zealand Medical Journal. 2017; 130(1452):54–62.
  29. Hibbert PD, Molloy CJ, Hooper TD, et al. The application of the Global Trigger Tool: a systematic review. International Journal for Quality in Health Care. 2016:1–10.