This study used a retrospective cohort design to determine associations between the nursing workload and adverse events among critically ill patients, with a focus on ICU readmissions and the occurrence of medication errors, for 3 consecutive months (February to April 2023). The sample was conveniently recruited from three different governmental hospitals in the UAE. These hospitals cover various specialties and sub-specialties in clinical services, including emergency, intensive care medicine and intensive cardiac care.

This study only included general ICUs from the participating hospitals; cardiac ICUs and cardiac care units were excluded because they have different workloads and capacities, and mainly cover cardiac cases. We reviewed the medical records for patients aged ≥18 years who were admitted to the ICU from February 2023 to April 2023 and who had experienced ICU readmission within the same hospitalisation period. The Shapiro-Wilk test was used to test normality in this study (p<.05). Based on this result, nonparametric tests, independent-samples Mann-Whitney U tests and correlation coefficients were used for the analyses. The sample size calculation using G* power software revealed that 270 participants would be sufficient to achieve 80% power with an alpha of 0.05.

The occurrence of medication errors was measured using the Categorizing Medication Errors Index, which is used to identify the harm based on the taxonomy for medication errors and index for categorising medication errors. This tool was developed by the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) in 2001. An algorithm was created to facilitate the error-index categorisation. We used this tool for the present study because it provided guidance (via a series of questions) for assigning the proper error type to an event. The tool contains nine medication error classifications based on severity and patient safety measures: circumstances/events that have the capacity to cause error (category A); an error occurred but did not reach the patient (an ‘error of omission’ does reach the patient) (category B); an error occurred that reached the patient but did not cause patient harm (category C); an error occurred that reached the patient and required monitoring to confirm that it resulted in no harm to the patient or required intervention to preclude harm (category D); an error occurred that may have contributed to/resulted in temporary harm to the patient and required intervention (category E); an error occurred that may have contributed to/resulted in temporary harm to the patient and required initial or prolonged hospitalisation (category F); an error occurred that may have contributed to or resulted in permanent patient harm (category G); an error occurred that required intervention necessary to sustain life (category H); and an error occurred that may have contributed to/resulted in the patient’s death (category I). These nine categories are divided into four classes: class I = near misses (category A); class II = error without harm (categories B, C and D); class III = error with harm (categories E, F, G and H); and class IV = error, death (category I) [35]. A previous study reported substantial agreement for this tool, with a Kappa score of 0.61 [35].

ICU departmental records were reviewed to identify patients who were readmitted to the ICU within the same hospitalisation period.

The primary researcher met the ICU heads of department and unit managers in the selected hospitals and explained the purpose of this study and was granted permission to access department records and patients’ medical records. ICU admission and discharge records were reviewed to identify patients who were admitted to the ICU from February to April 2023 and determine which patients were readmitted to the ICU within the same hospitalisation period. Patients’ electronic medical records were reviewed to obtain their demographic data and calculate their NAS based on the acuity level, comprehensive assessment, ongoing assessment tools, and healthcare providers’ notes (nurses and physicians). The NAS was calculated for the days of ICU admission and discharge for each patient to identify the nursing workload on these two days and allow us to determine the correlation between the nursing workload and adverse events. Moreover, medication error occurrences were obtained from ICU nurses who were auditing for medication compliance in the department, with the error type and category recorded based on the NCC MERP. All data were entered into a Microsoft Excel document without including patients’ identification details or medical record numbers.

This study was approved by the research ethics committees at Sharjah University (REC-22-12-18-S) and the Ministry of Health and Prevention (MOHAP/DXB-REC/F.M.A/No. 46/2023). Administrative permission was obtained from the Emirates Health Services and ICU heads of departments in the selected hospitals before accessing patients’ records. This study was completely anonymous. All collected data were coded and entered into a password-protected computer that only the research team could access.

Data were analysed using SPSS version 26. Descriptive statistics (frequency and percentage or mean and standard deviation [SD]) were used to describe patients’ sociodemographic characteristics. Descriptive statistics (mean and SD) were used to evaluate the nursing workload on ICU admission/discharge days and the prevalence of adverse events (ICU readmission and medication error occurrences). Independent-samples Mann-Whitney U tests and correlation coefficients were used to determine associations between nursing workload, clinical characteristics and adverse events.

In total, 270 electronic medical records from February to April 2023 were reviewed for this study. The majority of patients were male (71.5%) and 28.5% were female. Patients’ mean age was 54.13±16.60 years. Patients’ clinical characteristics are shown in Table 1. The most common reason for admission to the ICU was cardiovascular diseases (25.2%), including ST-segment elevation (8.1%), ischaemic heart disease (7%), post-cardiac arrest (6.7%) and non-ST-segment elevation (3.7%). The second most common reason for ICU admission was respiratory disorders (21.9%), including acute respiratory distress syndrome (12.6%), pulmonary oedema (5.2%) and chronic obstructive pulmonary disease (3.7%). Moreover, 15.6% of patients had neurological disorders, including stroke (7.4%), traumatic brain injury (6.3%) and seizure (2.2%). Other diagnoses such as gastrointestinal bleeding, meningioma and trauma cases accounted for 14.4% of patients. Around 12.2% of patients had medical disorders classified as sepsis (10%) and diabetic ketoacidosis (2.2%). Finally, 10.7% had nephrology disorders (end-stage renal disease: 9.3%, acute kidney injury: 1.5%).

The largest group of patients (54.8%, n=148) were admitted to the ICU from the emergency room, followed by inpatient wards (29.3%, n=79), the operating room (12.2%, n=33) and transfer from other institutions (3.7%, n=10). The mean length of ICU stay was 14.66 ± 11.47 days. The majority (63%, n=170) of patients required intubation and were under invasive mechanical ventilation; however, 17% (n=46) were treated with non-invasive mechanical ventilators. The mean mechanical ventilator duration was 8.87 ± 10.07 days. Furthermore, 16.3% (n=44) of patients were tracheotomised during their ICU stay. Most patients were discharged from the ICU during the morning shift (63.3%, n=171), and 36.7% (n=99) were discharged during the night shift.

For each patient, the NAS was calculated based on the patient’s condition on the days of ICU admission and discharge. The mean NAS was 72.61±14.80 on admission day and 52.61±15.38 on discharge day.

Table 2 presents the prevalence of adverse events (i.e. ICU readmission and occurrence of medication errors). We found that 23.7% (n=64) of patients were readmitted to the ICU within the same hospitalisation period, and 15.9% (n=43) had experienced medication errors. The most frequent medication error was wrong order (8.5%), followed by wrong time (5.2%) and wrong documentation (2.2%). The most common error category was category B (9.6%, n=26), followed by categories A (4.8%) and C (1.5%).

Independent-sample Mann-Whitney U tests were used to examine the relationships between adverse events and the nursing workload. There were significant associations (p=.000) between a patient’s ICU readmission and the nursing workload on ICU admission and discharge days. Moreover, there was a significant association (p=.006) between the occurrence of medication errors and nursing workload on the day of ICU admission. Furthermore, while assessing the relationships between the occurrence of ICU readmission and patient outcomes, significant effects were observed for the patient’s age, diagnosis, acuity score on the day of ICU discharge, isolation precautions on ICU admission and discharge days,

length of ICU stay, mechanical ventilation, duration of mechanical ventilation and tracheostomy during ICU stay. However, there were no significant relationships between medication errors and patient outcomes. Table 3 details the associations between the patient demographics and clinical characteristics and adverse events.

Moreover, we observed significant associations between the nursing workload on the days of ICU admission and discharge and patient outcomes, including acuity score at admission and discharge, isolation precautions on ICU discharge day, mechanical ventilation and tracheostomy during ICU stay. There was also a significant relationship between the nursing workload on the day of ICU discharge and the discharge shift. We examined the correlation coefficients to identify correlations between the nursing workload on the ICU admission and discharge days and length of ICU stay and mechanical ventilation duration. There was a moderate positive relationship between the nursing workload on ICU admission and discharge days and length of ICU stay (r=.336 and r=.323 for admission and discharge days, respectively). In addition, there was a moderate positive relationship between the nursing workload on ICU admission and discharge days and the duration of mechanical ventilation (r=.381 and r=.373 for admission and discharge days, respectively) (Table 4).

The nursing workload on the day of ICU admission was high. This may be because most patients were admitted with cardiovascular diseases, respiratory disorders or neurology diseases with underlying co-morbidities, all of which required comprehensive medical therapy and intensive nursing care. Similar findings were reported in previous studies [10, 13] that found most patients in the ICU needed long-term special medical and nursing care, such as ventilation support, central line insertion, observing pump machines and drains and administering medications. In addition to direct care responsibilities, indirect care and administrative assignments for newly admitted patients impact the nursing workload [36, 37]. Furthermore, a nurse may spend 2–6.1 hours per shift donning and doffing personal protective equipment, especially in the isolation room, which further increases their workload [38].

The ICU admission process is essential in setting the stage for patient care. Developing comprehensive admission guidelines that encompass thorough patient handovers, meticulous assessments and preparedness for potential emergency interventions is crucial. Effective communication and collaboration among ICU nurses, physicians and bed management teams will facilitate a streamlined admission process, allowing for a smoother transition and comprehensive care from the moment a patient enters the ICU [39].

Nursing workload management is therefore a critical aspect of healthcare practice, particularly in the demanding field of critical care nursing. To ensure optimal patient outcomes, it is essential to establish patient assignment strategies that go beyond simple nurse-to-patient ratios. Healthcare facilities should adopt an approach that considers the severity of a patient’s condition and the specific nursing activities required for each case. This personalised approach can significantly improve the allocation of nursing resources and ensure that patients receive an appropriate level of care. Furthermore, it is practical to use a validated tool such as the NAS to assess the nursing workload because it focuses on the time consumed for each nursing task (including direct and indirect care and clinical or administrative assignments). The NAS is used as a tool for arranging nurse-to-patient assignments in many different countries worldwide [11, 15, 40].

Moreover, establishing specialised units, such as geriatric departments that are dedicated to the unique needs of older patients, can ensure that healthcare providers are properly equipped to effectively address the challenges associated with chronic diseases and comorbidities. This approach will optimise patient outcomes and also promote a more holistic approach to healthcare delivery.

The majority of patients had high NAS scores on the day they were discharged from the ICU, which was consistent with previous findings [20, 21]. A possible reason for the high NAS at ICU discharge in our study may be that many patients were older and had a history of chronic conditions and comorbidities. A previous study [41] reported that patients who were discharged from the ICU with a high nursing workload included those who were bedridden or had a tracheostomy, chronic diseases and comorbidities. Moreover, the high workload on the day of ICU discharge may be explained by the number of responsibilities and tasks ICU nurses must complete before a patient is transferred to a ward, such as communicating with the bed management nurse to book a ward bed, follow-up for laboratory investigations and diagnostic procedures, preparing medication/documentation and handing over the patient to the transferred department. These tasks are time-consuming. It is important to note that if the nurse is also assigned to another patient, the completion of such tasks could affect patient safety and may compromise the quality of care [6, 16, 17].

The present study highlighted that nurses had a high workload on the days of ICU admission and discharge, and thus we found significant differences between the workload on ICU admission/discharge days and ICU readmission. Similarly, previous studies reported that patients discharged from the ICU with a high nursing workload had a greater probability of being readmitted to the ICU within the same hospitalisation than patients without a high nursing workload [14, 21]. The explanations for the findings in our study were that most ICU readmissions were older patients (mean age of readmitted patients: 60.67 years), patients who required invasive mechanical ventilation and those that were tracheotomised during their ICU stay, which was consistent with previous findings [20, 21]. However, another study [42] reported respiratory failure and circulatory instability were two major causes of ICU readmission, which was inconsistent with our study. Other factors that may be associated with ICU readmission reported in a previous study were comorbidities, length of ICU stay, discharge time (morning shift/night shift) and patients’ neurological function (including their Glasgow Coma Scale score) [41]. These findings suggested that to prevent ICU readmissions, healthcare facilities should undertake comprehensive risk assessments for patients who are at a high risk for readmission after discharge. Establishing specialised follow-up programmes for patients transferred to general wards or step-down units may also play a pivotal role in ensuring that patients’ needs are adequately met, and that the possibility of premature discharge is minimised [21].

The present study found a significant association between the nursing workload on the day of ICU admission and medication errors. Previous studies reported similar findings [6, 19, 28, 29]. Category B medication errors were the most common errors, with wrong order being the most prevalent error, including the wrong route of administration, missing titration and maximum dose in the order itself, the wrong time and wrong documentation. This was consistent with a previous study [25] that reported wrong order was the common cause of medication errors. However, another study [28] identified the wrong time as a frequent error. This could be explained by patients’ condition severity, possible procedures (e.g. central line insertion, intubation), the need to stabilise the patient and other indirect and administrative tasks. Poor communication, lack of knowledge, inexperience and fatigue were other commonly reported causes of medication errors [25].

These findings suggested that ICU nurses should be encouraged to adopt a double-check system for medication administration, which is consistent with the established ‘7 rights’ of medication administration. Regularly reviewing and validating patients’ medication charts can help identify errors and inconsistencies in orders, dosages, routes and other critical details. Furthermore, embracing technological solutions (e.g. automated dispensing cabinets and barcode medication administration systems) may substantially reduce the risk of medication errors and enhance patient safety [43]. Continuous education and training should be integral to nursing practice, especially in the context of patient safety and workload management. Addressing factors that contribute to medication errors, such as poor communication, lack of knowledge, inexperience and fatigue, could be achieved through targeted educational programmes and ongoing professional development opportunities.

In collaboration with healthcare policymakers, healthcare organisations should advocate for evidence-based staffing and workload policies that align with patient acuity and complexity. Conducting research to delve deeper into the intricate relationships between nursing workload and adverse events could inform policy decisions, and ultimately lead to improved patient outcomes and enhanced healthcare delivery. Finally, supporting the nursing workforce is paramount. Ensuring that nurses have access to regular breaks, mental health resources and opportunities for continuous professional development can help mitigate the impact of high workload on patient care and nurses’ well-being. By adopting these recommendations, healthcare organisations can cultivate an environment that prioritises patient safety, nurse satisfaction and overall quality of care in critical care settings.