Nurses' Perceptions of Workload Burdenin Pediatric Critical Care

Ruth M. Lebet, PhD, RN; Natalie R. Hasbani, MPH; Martha T. Sisko, RN; Michael S. D. Agus, MD; Vinay M. Nadkarni, MD; David Wypij, PhD; Martha A. Q. Curley, PhD, RN


Am J Crit Care. 2021;30(1):27-35. 

In This Article


Study Design and Procedure

We studied nursing workload as a prespecified secondary outcome of a 35-center randomized clinical trial testing the efficacy of a computerized GC protocol in critically ill pediatric patients with respiratory or cardiovascular failure (Heart and Lung Failure–Pediatric Insulin Titration Trial [HALF-PINT], NCT01565941). The HALF-PINT trial methods have been described in detail elsewhere.[17] Consented critically ill children aged 2 weeks to 17 years who were receiving vasoactive support to treat hypotension and/or invasive mechanical ventilation were randomized into 1 of 2 groups: a lower glucose target group (80-110 mg/dL) or a higher glucose target group (150-180 mg/dL).[17] As reported by Agus et al,[17] there were no significant differences between the participants randomized to the lower-target and higher-target groups in terms of demographic characteristics such as age, race, PRISM (Pediatric Risk of Mortality) III-12 score (severity of illness on PICU admission), or risk of death.

A bedside computerized algorithm incorporating continuous subcutaneous glucose and intermittent bedside glucose monitoring provided ongoing protocol instructions.[18] The HALF-PINT computerized decision-support system supported protocol implementation, theoretically decreasing cognitive burden[19] by completing calculations and recommending adjustments of the insulin infusion and providing a countdown to the next blood glucose measurement. The computer-based decision-support tool was incorporated into the nurse's workflow by its placement at the patient's bedside so that it was immediately available when a management decision had to be made. The tool required minimal data input and provided specific recommendations for next steps. Interprofessional compliance with the algorithm was high (97% compliance in the lower-target group, 99% in the higher-target group). The institutional review board at the primary site coordinated central ethics review for 10 study sites that established reliance relationships; at the remaining sites, oversight was provided by a local institutional review board.

All staff nurses working in the 35 participating PICUs who could be assigned to care for a HALF-PINT participant were invited to participate in the current study. Nurses were invited to complete a baseline survey at the start of the study, before caring for a HALF-PINT participant, and then an intervention survey after caring for a study participant. Completion of the surveys constituted participant consent. The 21-item baseline survey collected nurse demographic and clinical experience characteristics and nurses' perceptions about ICU nursing workload in general. The 36-item intervention survey queried nurses about their perceptions of workload related to the care of a patient on the HALF-PINT protocol. We randomized at the patient level, with a goal of obtaining 1 linked baseline and intervention survey for each enrolled participant. This was done to ensure that surveys were distributed across shifts and across the duration of the study and to reduce selection bias. Intervention surveys were distributed to nurses working on randomly selected shifts, selected based on the last digit of the participant identification number. The survey link was provided to the nurses during the course of their shift, and they were asked to complete the survey at the end of their shift. If a nurse was unable or declined to complete the intervention survey on the assigned shift, the nurse caring for the HALF-PINT participant on the next shift was asked to complete the survey. The project study coordinator tracked completed studies daily and notified the local study coordinator if an expected survey was not received, requesting that the survey be reassigned to the nurse on the next shift. If a participant came off of the protocol before the assigned shift, the nurse caring for the patient on the last eligible shift was asked to complete the survey. The 2 surveys were linked using the nurse-generated anonymous unique code (see Supplement 1—available online only at Nurses could be randomly selected to complete a survey for more than 1 participant, who could be in either the higher-target or lower-target group. The survey instruments are available from the corresponding author.


Nursing workload was assessed using 2 instruments, the Subjective Workload Assessment Technique (SWAT)[20] and the National Aeronautics and Space Administration–Task Load Index (NASA-TLX).[21–27] The SWAT employs a human factors framework and has been used to study the effect of workload in nursing, pharmacy, and medicine.[24,27,28] The SWAT is designed to measure 3 dimensions of workload: cognitive, time, and psychological stress burden. Cognitive burden is the amount of mental effort or concentration required for the complexity of tasks that affect work performance. Time burden refers to the presence or absence of spare time, interruptions, and task overlap that affect work performance. Psychological stress burden refers to the presence of elements such as confusion, risk, frustration, and anxiety associated with work that affect work performance. The NASA-TLX is a 6-dimension tool designed to obtain an overall perception of workload related to high-stress tasks. The 6 dimensions are cognitive demand (mental and perceptual activity required), physical demand (physical activity required), time pressure (rate or pace of activities), performance (individual success in completing assignments), effort (how hard one has to work to accomplish one's level of performance), and frustration (how irritated, stressed, or annoyed one feels during the day). Supplement 2 (available online only) provides additional information on these tools.

Both the NASA-TLX and SWAT require pretask rankings in which respondents are asked to identify contributors to workload that are most important to them. These items were placed in the baseline survey. The intervention survey contained the remaining SWAT and NASA-TLX questions and an adapted version of the Nine Equivalents of Nursing Manpower Use Score (NEMS).[29,30] The NEMS is a validated score that represents critical care bedside nurse workload. We adapted the NEMS by creating a single question asking nurses to select 1 NEMS item that was most similar to the work involved in caring for a HALF-PINT participant. Supplement 2 (available online only) provides additional information on the NEMS.

Statistical Analysis

Nurse demographic and clinical experience characteristics and nurses' perception of workload when caring for any ICU patient were described for the entire cohort using descriptive statistics. Pretask SWAT and NASA-TLX scores were used to compute overall SWAT and NASA-TLX scores. Differences in overall SWAT and NASA-TLX scores were compared between target groups using linear regression. Other survey outcome data were analyzed using linear, logistic, and ordinal or nominal logistic regression for continuous, binary, and ordered or unordered categorical variables, as appropriate. All regression analyses accounted for PICU site as a cluster variable using generalized estimating equations. All P values were 2-tailed and considered statistically significant at P less than .05. Analyses were conducted in SAS, version 9.4 (SAS Institute).