Effects of nursing workload on medication administration errors: A quantitative study

Abstract

BACKGROUND:

Medication administration errors by nurses form a high proportion of medical errors in medical institutions. Studies have shown that such errors are closely linked to nursing workload.

OBJECTIVE:

To quantitatively explore the effects of different types of nursing workloads on different medication administration errors.

METHOD:

Three medical institutions were selected as the objects of error data collection based on the following criteria: the medical institution experience in error data collection, the complete range of medical departments, and the institution size. Error cases were self-reported from all nurses in all medical departments. The relationship between the error types and nursing workload types were quantitatively examined using partial least squares and structural equation modeling.

RESULTS:

The study recorded 290 medication administration errors, and extracted four error types and nine nursing workload types. The workload type for each error type was also identified and the path coefficient was found to be between 0.087 to 0.416.

CONCLUSION:

This study confirmed the effect of workload on medication administration errors and determined a theoretical mechanism for this effect. Research results will provide the evidence for nursing managers to reduce workload and ensure quality in the nursing administration process.

Keywords

Medication administration errors nursing workload error types quantitative study

1 Introduction

Medication errors are one of the most common types of medical errors and are an important issue in patient safety [1, 2]. The National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP, 2021) defines a medication error 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.” In the UK, 237 million medication errors are reported on average every year, out of which 66 million are potentially clinically significant. Additionally, more than 1,700 people die in the UK every year due to such errors [3]. Habraken et al. investigated 53 cases involving medication errors that resulted in injuries or deaths and found that approximately 55% of the errors occurred at the stage of medication administration [4]. Meanwhile, Elliott et al. examined the errors at each stage of medication and found that 21.3% of errors occurred during the prescription stage, 1.4% during the transition stage, 15.9% during the pharmacist dispensing stage, 54.4% during the medication administration stage by nurses, and 7.0% at the supervision stage post administration [3]. Thus, nursing medication errors account for a relatively high proportion of the total medication errors and are called medication administration errors. A medication administration error is defined as “the deviation between the expected medication implementation contents and actual medication implementation contents in nursing” [5, 6].

Correct medication implementation involves administering the right medication to the right patient, at the right time, in the right dosage, and using the right route of delivery, among others [7, 8]. Medication administration is an important part of nurses’ daily tasks and takes up approximately 40% of their working hours [9]. Nursing work is characterized by high intensity, long duration, irregular working hours, and involves various types of tasks. Nurses often have to deal with high workloads, carry out multi-task coordination, and keep abreast of diverse and changeable work content involving direct interactions with patients [10, 11].

Nursing workload was defined as “nurses’ investment (direct and indirect) of time and care in patients, workplaces, and career development” [12 –14]. Campbell et al. have identified six types of nursing workloads: current patient workload, current medication workload, task workload, patient acuity, number of interruptions, and weekly working time [15]. Westley et al. discussed the influence of operational time load on nurses’ dosing errors, and proposed that the proportion of dosing errors occurring among nurses who worked more than 60 hours per week was 1% higher than that of nurses who worked less than 60 hours per week based on the interviews of 20 nurses [16]. Keers et al. proposed that workload was one of the important factors affecting the occurrence of nurse dosing errors [17]. Jin et al. proposed that workload was an important factor for non-observance behavior with operation standards in nursing administration, and explored the relationship between workload and non-observance errors in nursing [6]. Dominic et al., based on 721 questionnaires conducted for medical staff, proposed from a systematic perspective that Inadequate resourcing, Presenteeism and Unsupportive management are the main reasons for the high workload in medical institutions [18]. To summarize, previous studies have shown that the daily work of nurses are accompanied by high workloads, and there is a close relationship between medication administration errors and nurses’ workloads. However, in order to prevent medication errors from a workload perspective, the influence of different workloads on each medication error should be further discussed.

Medication administration errors directly affect patient safety, and nursing workloads are important factors that may induce medication administration errors. There are different types of medication administration errors, and several workload types based on the aspects of time as well as physical and cognitive conditions, making the relationships among them more complicated. From the error-prevention perspective, it is necessary to clarify which workload types affect which errors. Therefore, it is imperative to quantitatively explore the effects of different types of nursing workload on different medication administration errors by extracting errors and workload types using medication error case analysis.

2 Methods

2.1 Participating hospitals and nurses

Data on cases involving medication administration errors were collected from January 2019 to December 2020. Self-reports from nurses were collected from three general medical institutions (Hospital A: acute care type, 1,313 beds, China; Hospital B: acute care type, 1228 beds, China; Hospital C: acute care type, 892 beds, China). The reasons for choosing these three medical institutions are as follows. First, these three medical institutions have long experience in collecting medical errors. Second, they have comprehensive medical departments. Third, they are large medical institutions with a large number of patients.

2.2 Data collection

In most hospitals, errors made by nurses during the administration of medications to patients are recorded in medical error case reports and submitted to the medical safety management office. This study collected error cases from the medical safety management offices of three hospitals and was not directly conducted from the nursing medication administration site. The data obtained in this study are based on nurses’ self-report. The cases were collected for all medication administration errors by nurses in all departments of the three institutions.

The case reports of medication administration errors included the following information:

Expected and actual implementation of basic medication administration in terms of the type of medication administered, time, objective, dosage, route, and method, among others.

Expected and actual implementation at the stage of medication instruction, medication preparation, medication delivery, and observation after medication administration.

Other information about medical errors: self-analysis of errors, such as, how medical errors are detected, their causes, their operating characteristics (including workload conditions), and countermeasures that can be considered.

2.3 Data analysis and data inclusion/exclusion criteria

Each error case report was analyzed by two medical safety researchers and one medical safety administrator (both have more than 10 years of experience studying medical errors).

All cases of nursing medication administration errors were included in the error analysis. During the error analysis, errors attributed to workload were first extracted from all the reports. Next, each error was analyzed, and the error type was identified and workload was extracted. Cases that did not have an identifiable type of error and workload due to insufficient recording were excluded from the study.

2.4 Extraction of error types and workloads

Error types refer to the modes of error that existed in nursing medication administration. Extracting error types helps to systematically observe the trend in error occurrence, determine the main error types, and conduct targeted error prevention [19]. Considering that the errors covered by this study were human errors, the error types were extracted based on human cognition and behavior processing at the stages of perception, cognition, memory and action [20].

Workload refers to the amount of work per unit time and includes physical workload, cognitive workload, and time workload [12 –14]. To extract the workload, the standard and actual workflows of processes where errors are commonly reported were reviewed. Next, it was determined whether the nursing workload was high. Then, the specific workload types were identified.

Here is an illustration of the process through which error and workload types were extracted during the study:

Case: At one hospital, medication X was oftenadministered at a dose of 1 A. Therefore,during one night shift, a nurse adminis-tered medication X at a dose of 1 A withoutlooking at the prescription. However, theprescription had been changed and thenew dosage was supposed to be 2 A.

Under the standard workflow, before administering any medication, nurses need to confirm eight items: date, time, medication name, use, dosage, speed, route, and patient name; however, the dose of the medication cannot be accurately confirmed and is based on previous experience with administration.

The study used the following analysis process for this case:

Step 1: Error-type extraction

Step 1.1: Summary of error case content

The nurse should have administered medication X at a dose of 2A, but provided a dose of 1A instead.

Step 1.2: Identifying the error behavior

The dosage of medication X is usually 1A. Operating under this assumption, the nurse did not accurately confirm the dosage.

Step 1.3: Determination of error type

The error in this case was due to the nurse’s information cognition process and she did not recognize the relevant administration information. Therefore, the error type was determined as “Incomplete information cognition.”

Step 2: Workload-type extraction

Step 2.1: Combining standard workflow and actual workflow

Standard workflow: Nurses need to confirm the date, time, medication name, use, dosage, speed, route, and patient name before administration.

Actual workflow: The medication administration information was not confirmed according to the standard workflow. This resulted in the wrong dosage being administered.

Step 2.2: Workload analysis

The case involved a typical omission of cognitive information on part of the nurse to reduce workload. Under the standard workflow, nurses need to confirm eight items, hence, their cognitive workload is high. However, the nurse saw that medication X was a medicine commonly used in the past and its dose was usually 1A, so while administering the medication, she omitted to check the information on dosage based on predictive judgment.

Step 2.3: Determination of workload

The workload in this case fell under cognitive workload, which refers to processing a “large amount of information for cognition and judgment.”

2.5 Statistical analysis

Partial least squares (PLS) is a multivariate statistical data analysis method which combines the advantages of multivariate linear regression, principal component analysis, and canonical correlation analysis [21]. In this study, the PLS method was combined with structural equation modeling (PLS-SEM) to quantitatively express the relationship between nurses’ workload and medication administration errors. PLS-SEM was used because it: ding172 allows relationship modeling between multiple independent and multiple dependent variables; ding173 allows use of a small sample; and ding174 has no requirement for independence between independent variables.

SmartPLS 3.3.3 software was used to construct and calculate PLS-SEM, and the bootstrapping method was used for significance testing with a sub-sample size of 5,000. Goodness of fit (GoF) was used to verify the validity of the overall model. The calculation formula is shown in Equation (1): $G o F = \sqrt{\bar{communality} \times \bar{R^{2}}}$ (1) where, communality is the commonality of cross-validation of constructions, also known as common factor variance (obtained by the blindfolding algorithm in SmartPLS). R2 is the coefficient of determination, in which the calculation of average commonality is the weighted average of different commonalities, and its weight is the proportion of the number of variables of each latent variable.

3 Results

3.1 Extraction of medication administration errors

A total of 290 cases involving medical administration errors linked to nurses’ workload were extracted from the case reports collected from the three medical institutions.

Table 1 shows the total number of cases involving medication administration errors at each medical institution and the proportion of cases that are workload-related. The results show that the proportion of workload-related cases in medication administration errors is between 37.9% and 41.2%, whereas the overall ratio is 38.8%. It can, therefore, be said that the effect of workload on the results of clinical nursing work cannot be ignored.

Table 1
Distribution of medication administration errors

Hospitals Beds Medication administration errors (MAEs) Workload-related cases in MAEs Proportion of workload-related cases in MAEs

A 1313 326 124 38.0%

B 1228 235 89 37.9%

C 892 187 77 41.2%

Total 3322 748 290 100%

Hospitals	Beds	Medication administration errors (MAEs)	Workload-related cases in MAEs	Proportion of workload-related cases in MAEs
A	1313	326	124	38.0%
B	1228	235	89	37.9%
C	892	187	77	41.2%
Total	3322	748	290	100%

3.2 Types of errors and workloads

The results of our analysis of 290 case reports of medication administration errors are shown in Tables 2 and 3. Table 2 shows the four types of errors identified in the four stages of human cognition and behavior process. Table 3 shows nine types of workloads extracted from the three main types of workloads—time, physical, and cognitive—that are closely related to medication administration errors.

Table 2
Types of medication administration errors

Stage of error occurrence Error type Explanation Number and proportion

Perception Lack of perceptual information In the information perception state, missing the perception of some information. 65 (22.4%)

Cognition Incomplete information cognition In the information cognition stage, all the essential information is not considered properly or incorrect information is cognized. 89 (30.7%)

Memory Memory lapse In the memory stage of information, some work information is forgotten. 49 (16.9%)

Behavior Non-compliance with methods In the behavior stage, tasks are not conducted in accordance with the standard workflow but in another way considered more efficient by the nurses. 87 (30.0%)

Total 290 (100%)

Stage of error occurrence	Error type	Explanation	Number and proportion
Perception	Lack of perceptual information	In the information perception state, missing the perception of some information.	65 (22.4%)
Cognition	Incomplete information cognition	In the information cognition stage, all the essential information is not considered properly or incorrect information is cognized.	89 (30.7%)
Memory	Memory lapse	In the memory stage of information, some work information is forgotten.	49 (16.9%)
Behavior	Non-compliance with methods	In the behavior stage, tasks are not conducted in accordance with the standard workflow but in another way considered more efficient by the nurses.	87 (30.0%)
Total	290 (100%)

Note: The results retain decimals.

Table 3

Types of workloads associated with medication administration errors

	Workload type	Number and proportion	Total
Time	Excessive operation objects	75 (13.0%)	241 (41.8%)
	Excessive parallel operations	94 (16.3%)
	Different starting times of procedures	72 (12.5%)
Physical	Excessive operational materials	34 (5.9%)	134 (23.2%)
	Greater distance	64 (11.1%)
	Unsuitable methods	36 (6.2%)
Cognitive	Large amount of information for cognition and judgement	89 (15.4%)	202 (35.0%)
	Difficult information cognition	47 (8.2%)
	Frequent changes in information	66 (11.4%)
Total		577 (100%)

Note: The results retain decimals. A total of 577 workload data points were extracted from 290 medication administration errors because one error could be associated with multiple workloads.

According to Table 2, the stages at which errors occur in the process of medication administration are perception (2.2.4%), cognition (30.7%), memory (16.9%) and behavior (30.0%). Among the error types, incomplete information cognition accounts for the highest proportion of errors, followed by non-compliance with methods, lack of perceptual information and memory lapse. Therefore, when considering error countermeasures, the error types should be taken into account and the actual medical operation process should be combined to reduce the recurrence of the above error types by optimizing the operation method.

As shown in Table 3, nurses’ time workload (41.8%) and cognitive workload (35.0 %) are relatively high, while their physical workload (23.2 %) is relatively low. Under time workload, “excessive parallel operations” accounted for the highest proportion (16.3%) of errors; under cognitive workload, “large amount of information for cognition and judgment” accounted for the highest proportion (15.4%) of errors; and under physical workload, “greater distance” accounted for the highest proportion (11.1 %) of errors.

3.3 Quantitative relationship between error types and workloads

Table 4 shows the quantitative influence of nursing workload on medication administration errors based on PLS-SEM. All R2 values are greater than 0.3, with moderate and above explanation, which indicates that our model has good predictive ability. Usually, a path coefficient greater than 0.3 shows a strong causal relationship between the error type and workload. If the path coefficient is between 0.2 and 0.3, there is a moderate causal relationship; if the path coefficient is less than 0.2, there is a weak causal relationship.

Table 4
Effect of workload on medication administration errors

Endogenous variables R2 Path Path coefficient Bootstrapping (mean, STDEV, t-value)

Original sample Sample mean Standard deviation T-value

Lack of perceptual information (E1) 0.438 Unsuitable methods⟶E1 0.416^* 0.417 0.069 6.033 0.416

Large amount of information for cognition and judgement ⟶E1 0.239^* 0.236 0.052 4.615 0.239

Frequent changes in information ⟶E1 0.338^* 0.337 0.067 5.038 0.338

Incomplete information cognition (E2) 0.589 Excessive operation objects⟶E2 0.173^* 0.172 0.043 4.049 0.173

Excessive parallel operations⟶E2 0.285^* 0.286 0.049 5.798 0.285

Different starting times of procedures⟶E2 0.087^ 0.086 0.038 2.287 0.087

Large amount of information for cognition and judgement ⟶E2 0.235^* 0.233 0.048 4.848 0.235

Difficult information cognition⟶E2 0.158^* 0.158 0.043 3.710 0.158

Memory lapse (E3) 0.505 Excessive parallel operations⟶E3 0.235^* 0.233 0.048 4.848 0.235

Different starting times of procedures⟶E3 0.158^* 0.158 0.043 3.710 0.158

Excessive operational materials⟶E3 0.122^* 0.121 0.043 2.860 0.122

Frequent changes in information ⟶E3 0.394^* 0.392 0.055 7.147 0.394

Non-compliance with methods (E4) 0.557 Excessive operation objects⟶E4 0.285^* 0.286 0.049 5.798 0.285

Excessive parallel operations⟶E4 0.087^* 0.086 0.038 2.287 0.087

Greater distance⟶E4 0.235^* 0.233 0.048 4.848 0.235

Endogenous variables	R2	Path	Path coefficient	Bootstrapping (mean, STDEV, t-value)
Lack of perceptual information (E1)	0.438	Unsuitable methods⟶E1	0.416^***	0.417	0.069	6.033	0.416
		Large amount of information for cognition and judgement ⟶E1	0.239^***	0.236	0.052	4.615	0.239
		Frequent changes in information ⟶E1	0.338^***	0.337	0.067	5.038	0.338
Incomplete information cognition (E2)	0.589	Excessive operation objects⟶E2	0.173^***	0.172	0.043	4.049	0.173
		Excessive parallel operations⟶E2	0.285^*	0.286	0.049	5.798	0.285
		Different starting times of procedures⟶E2	0.087^**	0.086	0.038	2.287	0.087
		Large amount of information for cognition and judgement ⟶E2	0.235^***	0.233	0.048	4.848	0.235
		Difficult information cognition⟶E2	0.158^***	0.158	0.043	3.710	0.158
Memory lapse (E3)	0.505	Excessive parallel operations⟶E3	0.235^***	0.233	0.048	4.848	0.235
		Different starting times of procedures⟶E3	0.158^***	0.158	0.043	3.710	0.158
		Excessive operational materials⟶E3	0.122^***	0.121	0.043	2.860	0.122
		Frequent changes in information ⟶E3	0.394^***	0.392	0.055	7.147	0.394
Non-compliance with methods (E4)	0.557	Excessive operation objects⟶E4	0.285^***	0.286	0.049	5.798	0.285
		Excessive parallel operations⟶E4	0.087^***	0.086	0.038	2.287	0.087
		Greater distance⟶E4	0.235^***	0.233	0.048	4.848	0.235

Note: * t-value > 1.96 at p < 0.05, ** t-value > 2.58 at p < 0.01, *** t-value > 3.29 at p < 0.001; two-tailed tests.

Table 4 shows workload and error types that are significantly related. Taking the error type “lack of perceptual information” as an example, “unsuitable methods” has a strong causal relationship with the path coefficient of “lack of perceptual information” at 0.416; the path coefficient of “frequent changes in information” to “lack of perceptual information” is 0.338; the path coefficient of “large amount of information for cognition and judgement” to “lack of perceptual information” is 0.239. Figure 1 illustrates the causal relationships shown in Table 4.

Fig. 1

Structural model diagram. Bold lines represent strong causality, fine lines represent moderate causality, and dotted lines represent imaginary causality.

4 Discussion

The urgency, diversity, and complexity of clinical nursing work is reflected in the heavy workload of nurses. Specifically, clinical medication delivery is characterized by a heavy workload, diverse patient base, a wide variety of medications, and inconsistent administration times and dosages, which account for a high number of medical errors that can potentially affect patient safety. It is, therefore, imperative to explore the influence of nursing workload on medication administration errors.

Several studies have focused on the relationship between nursing workloads and medication administration errors. Keers et al. have pointed out the important role of nursing workloads in medical errors using actual error cases, but their study does not determine specific types of workloads or their impact on errors [17]. Fagerström et al. have measured the relationship between the daily workload of nurses and different types of patient safety incidents using experimental methods. Their study explains the effect of nurses’ workload on different types of medical errors but does not categorize workloads or explain the role of different workloads in different medical errors [11]. Jin et al. have extracted seven kinds of workloads related to nurses’ non-observance errors and discussed how increased workloads lead to non-observance. [6] Campbell et al. have discussed the impact of nursing workloads on medical errors using big data. They have divided workloads into six types from the perspective of the job object: patient count, medication count, task count, call light count, average sepsis score, and two-hour time periods [15]. However, their study also does not consider the influence of different workloads on different errors.

The current study not only extracted nine specific workload types and four medication error types, but also explored the impact of each workload on different medication error types using PLS-SEM. Therefore, the results of this study have the following unique characteristics:

First, the types of workloads extracted in this study are specific and match highly with actual medical operations because they have been extracted from actual error cases and therefore, have a strong practical significance. For example, the cognitive workload has three sub-types: large amount of information for cognition and judgment, difficult information cognition, and frequent changes in information. This shows that all medication administration information during the medication administration operations of nurses is based on the prescription information of doctors and pharmacists. Therefore, all medication administration information needs to be recognized (large amount of information for cognition and judgment). A large amount of information contains different handwriting styles of different doctors or pharmacists (difficult information cognition), and medication administration information can change at any time with a change in the patient’s condition (frequent changes in information).

Second, after determining the types of workload and error types in the process of nursing medication administration, this study used the PLS-SEM quantitative model to determine how the workloads affected the error types. There are various types and inducing factors in medication administration errors, but the current related research has not paid enough attention to the relationship between them, leading to the lack of pertinence in the prevention of medication administration errors. The main reason that can be taken into account is that previous relevant studies were mainly conducted in the form of questionnaires or interviews, and it is difficult for research data to support quantitative discussion of their relationship [15 –18]. In this study, based on the collection of actual medication error cases, the quantitative data of error types and workload were sorted out to study the relationship between them. It expressed the types of errors caused by each workload and showed the effect of workload on each error type. The research results not only explained the correlation between them from the academic point of view, but also provided ideas and basis for practical error prevention.

4.1 Limitations

There are some limitations to this study. First of all, the object of this study is medication administration errors in medical treatment, so the relevant research results may be difficult to apply to medical errors in other fields. Secondly, although the relationship between workload and medication error types were presented in this study, specific countermeasures based on the relationship were not discussed. Finally, the research results are based on the data of 3 medical institutions, and the number of medical institutions and error cases should be further increased.

5 Conclusions

Four types of errors were summarized in this study based on the human information processing process in actual medication administration error cases, and nine types of workloads in the error cases were extracted from the three workloads of time, physical, and cognition. PLS-SEM was used to quantitatively study the causal relationship between workload and medication administration errors. The results of this study show that external workload is indeed one of the causes of nursing medication administration errors. Based on our analysis, we suggest that hospital managers pay attention to the workload of nurses and start implementing various methods to deal with the situation.

Conflict of interest

The authors declare that they have no conflict of interest to disclose.

Funding

This work was supported by the National Natural Science Foundation of China (Grant nos 72171042 and 71701039) and the Fundamental Research Funds for the Central Universities (Grant no. N2106007).

Footnotes

Acknowledgments

This research was carried out with the support and cooperation of three medical institutions in China and has received valuable views and suggestions from experts in the medical care field. The authors would like to express their sincere gratitude to all experts and scholars for their help in completing this thesis. They would also like to thank Editage (www.editage.cn) for English language editing. Furthermore, they thank the editor and anonymous reviewers for their valuable comments and advice.

References

Khanikosarkhizi

, Mohammad

ASR

, Mehdi

. Safety climate and related factors in rehabilitation nurses of hospitals in Iran. Work. 2021;68(1):189–96.

Suclupe

, Martinez

, Mancebo

, Font

, Castillo

, Viñolas

, Robleda

. Medication errors in prescription and administration in critically ill patients. J Adv Nurs. 2020;76(5):1192–200.

Elliott

, Camacho

, Jankovic

, Sculpher

, Faria

. Economic analysis of the prevalence and clinical and economic burden of medication error in England. BMJ Qual Saf. 2021;30(2):96–105.

Habraken

, Schaaf

. If only...: failed, missed and absent error recovery opportunities in medication errors. Qual Saf Health Care. 2010;19(1):37–41.

Jin

, Qu

, Munechika

, Sano

, Kajihara

, Duffy

, et al. Applying intelligent algorithms to automate the identification of error factors. J Patient Saf. 2021;17(8):e918–e928.

Jin

, Chen

, Munechika

, Sano

, Kajihara

. The effect of workload on nurses’ non-observance errors in medication administration processes: A cross-sectional study. Int J Nurs Pract. 2018;24(5):e12679.

Ahmed

, Azer

, Abd

. Effect of In-service Training Program on Minimizing Nursing Intravenous Medication Process Errors by Syringe through Intravenous Catheter. Am J Nurs. 2019;8(4):153–62.

Fujihara

, Goto

, Higashino

, Nakamura

, Tanaka

, Sunaga

, Sasaki

. The impact of lesion location on medication self-management ability in patients with cerebrovascular disease. Japa J Comp Reh. 2020;11:21–7.

Muzio

, Dionisi

, Simone

, Cianfrocca

, Muzio

, Fabbian

, Giannetta

. Can nurses’ shift work jeopardize the patient safety? A systematic review. Eur Rev Med Pharmacol Sci. 2019;23(10):4507–19.

10.

Simões

, Sa

, Simões

, Oliveira

, Santos

, Mateus

, Magalhães

, Martins

. Nursing workload assessment in an intensive care unit: a 5-year retrospective analysis. J Clin Nurs. 2021;30(3–4):528–40.

11.

Fagerström

, Kinnunen

, Saarela

. Nursing workload, patient safety incidents and mortality: an observational study from Finland. BMJ Open. 2018;8(4):e16367.

12.

Ebrahimi

, Jafarjalal

, Lotfolahzadeh

, Kharghani

. The effect of workload on nurses’ quality of life with moderating perceived social support during the COVID-19 pandemic. Work. 2021;70(2):347–54.

13.

Chomem

, Motter

. Analysis of physical and physiological workloads of nursing in the surgical center. Work. 2021;68(2):425–35.

14.

Mohammed

, Alghamdi

. Nursing workload: a concept analysis. J Nurs Manage. 2016;24(4):449–57.

15.

Campbell

, Harlan

, Campbell

, Mulekar

, Wang

. Nurse’s Achilles heel: Using big data to determine workload factors that impact near misses. J Nurs Scholarsh. 2021;53(3):333–42.

16.

Westley

, Peterson

, Fort

, Burton

, List

. Impact of nurse’s worked hours on medication administration near-miss error alerts. Chronobiol Int. 2020;37(9–10):1373–6.

17.

Keers

, Williams

, Cooke

, Ashcroft

. Understanding the causes of intravenous medication administration errors in hospitals: a qualitative critical incident study. BMJ Open. 2015;5(3):e5948.

18.

Dominic

, Dipesh

, Amandip

. ‘It’s like juggling fire daily’: Well-being, workload and burnout in the British NHS-A survey of 721 physicians. Work. 2021;70(2):395–403.

19.

Walsh

, Hansen

, Sahm

, Kearney

, Doherty

, Bradley

. Economic impact of medication error: a systematic review. Pharmacoepidem Dr S. 2017;26(5):481–97.

20.

John

. Cognitive Psychology and Its Implications. 9th ed. Worth Publishers; 2020.

21.

Kumar

. Partial Least Square (PLS) Analysis. Resonance. 2021;26(3):429–42.