High-risk medication errors: Insight from the UK National Reporting and learning system

Abstract

Background

Ensuring patient safety is of paramount importance in healthcare systems. Rising concerns about medical errors in the UK have necessitated a greater focus on studying the nature of such errors, particularly those involving high-risk medications.

Objectives

To conduct a retrospective analysis of incidents related to patient safety in the UK based on data from the National Rporting and Learning System (NRLS).

Methods

This study was conducted based on a review of the National Reporting and Learning System (NRLS) patient safety reports published between January 1, 2015, and December 31, 2015. NHS Improvement provides details regarding incidents following approval using a data-sharing agreement. In total, 1500 incidents were analszed and equally divided among the three categories of high-risk drugs: opioids, insulin, and anticoagulants. Excel® features and deductive reasoning (thematic analysis) were used for data analysis.

Results

The results showed that the insulin category had both the highest risk and most errors compared with anticoagulants and opioids. These errors primarily result from issues related to administering, prescribing, and dispensing the drugs. Inadequate drug checks, communication difficulties among staff and patients, and high staff workloads are often linked to these errors.

Conclusion

This study confirms that the NRLS database is a valuable source of data, and the suggestions put forth, based on these results, could contribute to the formulation of measures that diminish the occurrence of errors related to high-risk drugs in healthcare settings. Information technology should enhance medication safety by tracking the process of medication use.

Impact of findings on practice statements

• •With the aid of the NRLS database, this study is the first to highlight the potential errors associated with the use of opioids, insulins, and anticoagulants, which should consequently enhance safety measures, increase training, and revise guidelines for administering, prescribing, and dispensing insulin.
• •Continuous analysis and learning from this reporting system can improve the healthcare delivery process and prevent medication error.
• •Emphasizing the nature, types, and causes of safety incidents with these medications can promote better healthcare staff communication, workload management, and continuous evaluation of patient safety.

1. Introduction

According to the World Health Organization (WHO), pharmacovigilance is a science and activity related to the detection, assessment, understanding, and prevention of drug-related problems, including adverse drug events (ADE).¹ They play a crucial role in promoting patient safety by preventing further issues associated with medication use. In the United Kingdom, the National Health Services (NHS) and Medications and Healthcare Products Regulatory Agency (MHRA) developed a National Report Learning System (NRLS) to report and monitor pharmacovigilance issues, including medication errors and adverse drug reaction reports.² The NRLS database has been standardized across all NHS trustsand designed to be accessible to all health provider for voluntary reporting.

The prescription process of medications is complex and risky. Owing to a rising patient population, the NHS has observed anincrease in the number of prescriptions issued daily.³ This increase can lead to a corresponding increase in the potential for medication errors and a huge economic burden, despite advancements in medical technology and treatment strategies. The cost of drug-related injury treatments is estimated to be at least $3.5 billion annually. Besides,that medication errors are estimated to incur $77 billion annually in morbidity and mortality costs.⁴

A medication error is defined as “any preventable event that may cause or lead to inappropriate medication use or patient harm while the medication is in the control of the healthcare professional, patient, or consumer”. Medication errors are categorized into two main categories: commission and omission.⁴ These errors can occur throughout the medication use process, including during prescription, dispensing, administration, and monitoring. Research indicates that a significantly small number of dispensing errors have been found and identified at the final stage in UK hospital pharmacies.⁵

High-risk medications can be defined as those that have a high potential for causing severe injury or harm if misused or administered incorrectly.⁶ These types of medications do not necessarily have significant error rate differences from other medications; however, when misused or used in error, they cause severe harm or death. Hospitals identify high-risk medications based on published evidence (for example The Institute for Safe Medication Practices (ISMP) or International Medication Safety Network (IMSN) of their potential to cause harm. These medications are then labeled as “high alert” and accompanied by specific policies and guidelines to aid clinicians in safe prescribing, dispensing, and administration. Opioids, insulin, and anticoagulants are considered the most frequent medication errors due to their high potential for errors, particularly those related to dosing. Therefore, these medications require precise and frequent dosage calculations.g.,d For example, relying solely on weight is insufficient and can increase error risk.⁷ This study aimed to investigate the nature, types and causative factorsaexample,nts of opioids, insulin, and anticoagulants medication errors by analyzing the risk of error in reported incidents involving these high-risk drug categories through the NRLS.

2. Methods

This study employed a mixed-methods research design that integrates both quantitative and qualitative approaches to data collection and analysis.

2.1. Data source

The NRLS database was usedas the primary data source. Between May and June 2016, the database was queried for incident reports published between January and December 2015. A comprehensive search strategy was employed utilizing both generic and trade names of medications within three high-risk drug categories: opioids, insulin, and anticoagulants. This approach ensured that all relevant patient incident reports associated with these drug classes were captured. Four key aspects of the incident reports were collected: (a) incident description, (b) apparent causative factors, (c) medication process involved, and (d) categorization of medication errors.

Owing to the extensive volume of retrieved reports (n = 51,994), a random sampling technique was used to select a representative subset for analysis. Following established guidelines recommending MS Excel ® for efficient randomization of samples exceeding 100 items, the RAND functionality tool in Excel was employed.^7,8 This process yielded a random sample of 500 reports for each of the three drug categories, resulting in a total sample size of 1500 reports.

2.2. Data analysis

Prior to the primary analysis, a pilot study was conducted to evaluate the robustness and reliability of the selected data-collection methodologies. A random subsample of 150 NRLS records (50 per high-risk drug category) was drawn from a larger dataset and subjected to a preliminary analysis.

Following the successful completion of data collection, quantitative data were compiled and analyzed using Excel software. This facilitated the efficient management and analysis of the numerical data in alignment with the objectives of the study. Qualitative data extracted from the incident report narratives were subjected to a rigorous deductive thematic analysis. This approach enabled the systematic identification of specific insights and development of logical conclusions based on broader themes and contextual information.

2.3. Inclusion criteria

Incident reports published between January and December 2015 formed the basis of the present study. To be included in the analysis, reports had to meet the following criteria: (a) involve the administration of anticoagulants, opioids, or insulin (categorized as high-risk medications) and (b) originate from primary, secondary, or tertiary care settings. The study population was comprised of patients of all ages and sexes.

2.4. Exclusion criteria

Incident reports were excluded from the analysis if they were associated with the following: (a) drugs other than anticoagulants, opioids, and insulin; or (b) codeine phosphate at a dosage of 60 mg or less for non-pain relief indications.

2.5. Ethical considerations

Following ethical approval from the University of Hertfordshire's Pharmacy College and a data-sharing agreement obtained in April 2016, data collection commenced. The data were stored securely on a password-encrypted personal drive. All research procedures, including quantitative and qualitative analyses, were conducted using a personal computer.

3. Results

As shown in Fig. 1, this study employed a four-phase approach: Phase 1: Initial Data Categorization: From the complete set of 51,994 NRLS incident reports, three primary medication categories were identified: insulin (n = 12,188; 23.44 %), anticoagulants (n = 17,910; 34.45 %), and opioid drugs (n = 21,896; 42.11 %). Phase 2: Random Sample Selection: A random sample of 550 reports was drawn from the NRLS database, proportionally representing each of the three medication categories. Phase 3: Pilot Study and Methodological Validation: A pilot study was conducted using a randomly selected subset of 50 reports from the phase 2 sample. This pilot phase aimed to ensure the validity and reliability of the data analysis procedure. The findings from the pilot study were compared to a separate analysis of the remaining 500 reports to confirm methodological rigor. Phase 4: Primary Data Analysis: The final phase of analysis was conducted on the remaining 500 reports per medication category (1500 reports in total), excluding those used in the pilot study.

Fig. 1.

Diagram of the data structure for randomization and analysis.

3.1. Types of medication error among the 51,994 reports

The highest percentages of medication errors were seen in administration (25,166 errors, 48.40 %), followed by prescription errors (11,564 errors, 22.24 %) and the preparation of medications in all locations/dispensing in a pharmacy (4574 errors, 8.80 %), as shown in Table 1.

Table 1.

Types of medication error for the 51,994 reports.

3.2. Sub-types of medication error for the 51,994 reports

The highest sub-types of medication errors among the three main categories were omitted medication/ingredients [insulin: 32.14 % (3917 reports); anticoagulants: 32 % (5732 reports); and opioid drugs: 11.42 % (2500 reports]), as shown in Table 2.

Table 2.

The highest sub-types of medication error from the 51,994 reports.

3.3. Medication errors involved among studied categories of medications

The main drugs involved in medication errors in the studied medication categories are reported in Table 3.

Table 3.

The main drugs involved in medication errors within the studied categories of medication.

3.4. Degree of harm for the studied categories of medications (n = 51,994)

The degree of ‘No Harm’ was the highest incident and the degree of ‘Death’ was the lowest incident level in the general studied categories of medication (Table 4).

Table 4.

Degree of harm for the three studied categories of medication (n = 51,994).

3.5. Patients' age range for the three studied categories of medication (n = 51,994)

The highest number of reported incidents occurred in patients aged >66 years. The details are presented in Table 5.

Table 5.

Patients' age range for the three studied categories of medication (n = 51,994).

3.6. Care setting locations of medication incidents for the three studied drugs

The top care settings of those incidents that occurred among the main three categories of medication were as follows: acute/general hospital, 76.13 % (39,583 reports); community nursing, medical, and therapy services (including community hospitals), 15.97 % (8303 reports); and mental health service, 4.52 % (2350 reports) (Table 6).

Table 6.

Care settings of incidents occurrence for the three categories of medication (n = 51,994).

The NRLS search output concerned a large number of incidents (51,994 reports) among the three main drug classes, namely Insulins, Anticoagulants and Opioid Drugs. Therefore, the study used only 500 reports, which were selected randomly according to the randomization method explained at the beginning, for further and specific data analysis from each category.

3.7. Types of medication error for each of the studied drug categories

The types of medication errors found among the randomization sample (1500 reports), which were taken from the original database (51,994 reports) by using PivotTable in Excel for the three categories of medication (Insulins, Anticoagulants and Opioids), are presented in Table 7. Following the application of the exclusion criteria, a total of 1330 incident reports were classified according to the type of medication error identified as follows: 48.4 % of all errors for all three categories were attributed to incorrect administration/supply of medication, and the number of errors related to insulin was the highest at 50.6 %, followed by opioid drugs (48.4 %) and anticoagulants (46.2 %). The second most common type of medication error was prescribed medication (24.33 %), with errors related to anticoagulants (31.80 %) being greater than those related to insulin (23.40 %) and opioids (17.80 %).

Table 7.

Types of medication errors for the main drug categories (n = 1330 reports).

In addition, administration errors were the most common type of insulin-related medication errors (50.60 %). The most common type of medication error related to opioid drugs was administration errors (48.40 %), followed by prescribing errors (17.80 %), and preparation/dispensing errors (11.20 %).

3.8. Sub-types of medication errors for each of the studied drug categories

Omitted medication/ingredient, 26.33 % (395 reports) followed by wrong/unclear dose or strength, and 13.47 % (202 reports) had the highest number of subtypes of medication errors (Table 8).

Table 8.

The highest sub-types of medication errors (n = 1192 reports).

3.9. Reported degree of harm for the three studied categories of medication

There were three cases of severe harm, while 13.31 % (180 patients) reported low harm,the highest number of people who reported low harm was in the Insulins category (Table 9).

Table 9.

Degree of harm for the three categories of studied medication (n = 1352).

3.10. Nature of errors for the three categories of medication

The most common types of errors for the three categories of medications were not adequately checking (task), 26.73 % (401); high workload (work environment), 10.67 % (160 reports); and communication problems (team), 8.40 % (126 reports) (Table 10).

Table 10.

Nature of errors across the three categories of medications (n = 1223 reports).

3.11. Quality of learning from the reports

used the incidents rating tools on the range of items included in each tool to judge each incidents to be of “good,” “fair,” or “poor” quality. The quality of learning gleaned across the three main categories of medication was poor, 61.01 % (834 reports), and many of these poor reports belonged to anticoagulants, 75.82 % (345 reports), followed by opioid drugs, and 67.28 % (292 reports), as shown in Table 11.

Table 11.

Quality of learning for the three categories of medication (n = 1367 reports).

3.12. Thematic analysis

Thematic analysis was conducted to identify the type of error based on the accidental causational model of reason across the three main medication categories. After coding the nature of all errors, refining the codes, and grouping them into categories, the following themes were generated and defined: There were two major themes in the Insulin, anticoagulant and opioid drug categories: system factor errors, which represented individual factors, design of equipment and supplies; and human factor errors, which represented individual, team and patient factors.

The first theme established is Individual Factor, such as ‘no adequate checking, as illustrated in the following example:

“Nurse not adequately checking and review TTO: - Insulin Missing – and other medication complete on TTO.”

The second theme is ‘Team Factor,’ as illustrated by the following example which relates to ‘protocols not followed’:

“Patient admitted with diabetic ketoacidosis. Sliding scale insulin not recommenced in ED as per plan by ED Dr.”

The third theme is ‘Design of Equipment and Supplies,’ and the following example illustrates issues of ‘product’ and ‘packaging storage’:

“Pharmacist dispensing patient's medication Humalin mix 30 instead Humalin mix 30 kiwipen because same packaging product.”

The fourth theme is the ‘Patient Factor,’ and the following example illustrates the issues of ‘task’ and ‘patient refusal’:

“Routine daily visit patient to administer prescribed daily dose of Insulin, patient refused to administered without any realistic reason.”

4. Discussion

High-risk medications are associated with an increase in the incidence of medical harm. This research aims to conduct a retrospective analysis of incidents regarding patient safety in the UK based on data from the National Report Learning System (NRLS). The research presented here was secondary research conducted based on a review of NRLS patient safety reports published from January to December 2015. It investigated the increasing safety risks associated with incidents involving medication and discussed emerging themes concerning such medication incidents.

Insulins represented the highest number of incidents that reported ‘low harm’ (18.03 %), followed by anticoagulants (11.45 %) and opioid drugs (9.98 %). The degree of ‘moderate harm’ attributed to insulin-related incidents was the highest at 2.52 % (12 patients) in comparison to anticoagulants (1.54 %) and opioid drugs (1.19 %). Recently, in the UK, safety issues concerning the use of insulin raised significant concerns, especially in view of the fact that some of the many delivery devices are a cause of confusion for patients, and further confusion arises concerning overdose adjustment of medication.⁹ Similarly, in June 2010, a report published by the National Patient Safety Agency (NPSA) in the UK raised grave concerns regarding the issue of safety associated with the administration and prescription of types of insulin by independent health institutes and the NHS and called for urgent action to reduce any further medication errors involving insulin.¹⁰ This alarming indicator of insulin-related medication errors requires further and deeper studies using verified data from the NRLS. The findings of this current study are alarming and should be of concern to all healthcare providers, such as consultant doctors, GPs, pharmacists, nurses, health policymakers, and healthcare management, and should prompt further action targeted at the minimization of risk and harm associated with medication errors, especially with regard to insulin-related cases.

Medication errors are defined as any preventable event that may cause or lead to inappropriate medication use or patient harm while the medication is in the control of the healthcare professional, patient, or consumer. Such events may be related to professional practices, healthcare products, procedures, and systems, including prescribing, order communication, product labeling, packaging, nomenclature, compounding, dispensing, distribution, administration, education, monitoring, and use.¹¹ The findings of the current study indicate that administration, prescribing, and dispensing errors from a total of 11 error types were the major causes of risk and harm in relation to the three categories of medication (insulin, anticoagulants, and opioid drugs). In the current study, half of the medication errors were related to drug administration, with the most common medication errors being related to insulin drugs (50.60 %), followed by opioid drugs (48.40 %), and anticoagulant drugs (46.20 %). Similarly, a review of relevant literature revealed that half of the total number of medication error reports (n = 12,552) represented hospital medication administration errors (MAE) with regard to young patients.¹² This indicates a high level of risk when half of the medication errors are related to drug administration, involving human and system errors. Moreover, concern was raised by Kelly et al. (2011) reported that administration errors in hospitals were between 3.0 and 8.0 %, and that prescribing errors in hospitals had reached 7 %.¹² When combined with high-risk drugs, medication errors can be severe and may even endanger the lives of patients. This is especially true in the case of medication errors related to the administration of drugs, as outlined by the ISMP, which are considered high-risk.

In this current study, it was found that the reported degree of harm for the three categories of medication showed no incidences of death occurring in the three categories. However, Three patients were reported to have experienced ‘severe harm,’ and there were 180 patients (13.31 %) who were reported to have experienced ‘low harm’; the highest number of cases which reported ‘low harm’ was in the insulin category. The degree of ‘moderate harm’ was also the highest for incidents involving insulin 2.52 % (12 patients), while anticoagulants accounted for 1.54 % (7 patients), and opioid drugs 1.19 % (5 patients). Between October 2011 and March 2012, 612,414 patient safety incidents were reported to the NRLS in England. Of these, a moderate level of harm occurred in 6 % of incidents, while a severe level of harm or death occurred in 1 % of incidents (n = 5235).^13,14

This study found that, in general, the quality of learning for the three main categories of medications was poor (61.01 %), and many of the poor reports belonged to the categories of anticoagulants (75.82 %) and opioid drugs (67.28 %). In contrast, adequate quality of learning was lower for the three main categories (3.80 %), and most of these adequate reports belonged to the insulin category (9.41 %), while the other categories accounted for 1.57 %. Based on the findings of the current research presented above, there remains significantly low quality in the reporting of incidents in the NHS in some medication categories, such as anticoagulants and opioid drugs, which prompts a strong recommendation here to health policymakers in the NHS to deal with this concern expeditiously. Similar studies have raised this concern, such as Donaldson et al. (2014), who examined NRLS reports on deaths due to care errors and categorized them according to the dimensions of systemic failure.¹⁵

Thematic analysis was used as part of this study in order to identify the nature of errors in relation to the main three medication categories: insulin, anticoagulants, and opioid drugs. This was achieved by reviewing and reading all reports/incidents not excluded by the research criteria, and subsequently coding all the nature of error types (See Appendix 1), refining the codes, and grouping them into categories. Based on these categories, major themes and sub-themes were generated and defined. Researchers have identified several factors contributing to patient safety incidents. There were factors that contributed to patient safety incidents, which were similar among insulin, anticoagulant, and opioid drug cases. Therefore, only the factors extracted from insulin-related data are presented here. Most insulin-related errors are caused by human mistakes, such as lapse of concentration and forgetfulness.16, 17

This study identified two major themes across the three main categories of medication: insulin, anticoagulants, and opioid drugs. These were extracted from the above themes: (a) system factor errors, which represent the individual factor, design of equipment, and supplies, and (b) human factor errors, which represent the individual, team, and patient factors. High-risk medications do not have significant error differences compared to other medications; however, when applied incorrectly, the potential for harm to the patient is considerably higher than in other cases.¹⁸

5. Strengths and limitations

The strength of the current research lies in the analysis of a one-year period of NRLS data for 2015, using a pilot study to ensure reliability and validity, and the use of thematic analysis to enrich the findings, particularly regarding the nature of errors and exploring their contributory factors to patient safety. The major limitation of the current study is its small sample size, which restricts the scale of presenting the phenomena of the research.

During the course of this research, various challenges and limitations must be acknowledged and addressed. Among them are the following:

• -Limitations associated with this study include the randomization process, which reduced the number of incident reports, as well as the data, which were not representative. In addition, as part of this study, it was not feasible to examine the severity of incidents or cases resulting in death owing to time restrictions. The sample size used in this study was small, and the period of the study was one year, from January 2015 to December 2015; however, if the study period was extended to six months, then the random sample could be more representative.
• -Some incidents were repeated, referred to as duplicate incidents, and were therefore excluded from the analysis.
• -Some incidents lacked, or had little, description; they were classified as “others” and also excluded from the analysis.
• -The description of some incidents mentioned Noval Oral Anticoagulant (NOAC); however, a search did not produce any drugs in this group because the term was entered in its acronym form.
• -Sometimes, an incident report was written by an individual who wished to complain about the procedure. Such incidents are referred to as ‘call for concern’ cases. This type of incident report was typically written by someone not satisfied with the hospital policy and who wished to change or amend the existing procedure.

Any errors related to medical devices, such as insulin syringes, should not be treated as medication errors. Therefore, the interpretation and analysis of incidents require researchers to have good experience and knowledge.

Viewing incident reports in the NHS database posed a challenge in terms of understanding the abbreviations commonly used in the reporting of incidents. Therefore, it was necessary to consult the terms used in the NHS Abbreviation List to fully understand the content of the incident reports as part of the analysis.

• -Dispensing errors can occur when a doctor prescribes warfarin (3 mg) and a nurse administers a different drug, or when a pharmacist dispenses NovoMix instead of NovoRapid insulin. Additionally, some incidents contained more than one error. When both dispensing errors and administration errors were found in the same incident report, such cases were categorized as ‘multi-process errors.’

In some cases, multiple themes were identified in the same incident. One incident included three themes that were relevant to the analysis. Consequently, researchers must carefully analyze incidents in the NHS database.

Deductive reasoning (thematic analysis) was used to identify themes from the original incident reports. Although this approach is interesting, using it for the first time within the limited timeframe of this study proved to be a challenge.

Based on the findings of the current study, it would be useful to review the incidence of medication errors every year using various methods to evaluate patient safety, such as questionnaires, interviews, observation, and case studies. In addition, the review and evaluation of patient safety and medication errors involved several patients at various levels. MHRA and the NHS strongly recommended regular evaluation of national medication safety in the UK; they suggest that the staff of health providers should be encouraged to engage in a systemic way with medication incidents. Further research can be specifically targeted at investigating the major medication errors that significantly affect patient safety, such as the administration, prescription, and dispensing of medication in the NHS compared with the private health sector in the UK.

6. Recommendation

The results obtained from these studies are essential for practice and policy, as they highlight the high rate and type of medication errors. Furthermore, the factors of medication errors and the strategies to mitigate them include the need to improve the education and training of healthcare providers, developing local clinical guidelines, involving pharmacists in leading diabetic, opioids, and anticoagulation clinics, and improving communication and adherence to be the best evidence-based strategies associated with diabetes, opioids, and anticoagulation drugs.

7. Conclusion

In summary, insulin produced more medication errors than anticoagulants and opioids. Moreover, the greatest number of errors could be attributed to three factors: administration of drugs, prescription of drugs, and dispensing of drugs. These three factors were related to the other three components that determined the nature of medication errors, such as lack of adequate checking of drugs, communication difficulties among staff and between staff and patients', and high staff workload.Furthermore, the overall quality of learning in relation to the three main categories of medications was found to be poor, and adequate quality of learning was very low for all three main categories, and most of these adequate reports related to insulin. Health information technology is a crucial tool for enhancing patient safety by tracking the medication-use process. In addition, improving communication skills leads to easy tracking of medication errors.

Funding

This research was funded by the Researchers Supporting Project number (RSPD2024R778), King Saud University, Riyadh, Saudi Arabia.

CRediT authorship contribution statement

Abdulrhman Alrowily: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Khalid Alfaraidy: Writing – review & editing. Saleh Almutairi: Writing – review & editing. Abdullah Alamri: Writing – review & editing. Wejdan Alrowily: Writing – review & editing, Conceptualization. Mohammed Abutaleb: Writing – review & editing, Formal analysis, Conceptualization. Mohammad Zaitoun: Writing – review & editing, Validation, Conceptualization. Waddad Sarawi: Writing – review & editing, Funding acquisition, Conceptualization. Mashael Aljead: Writing – review & editing, Conceptualization.

Declaration of competing interest

The authors declare no conflicts of interest.

Appendix A. Supplementary data

Supplementary material