Measuring workload with a modified NASA-TLX following Electronic Medication Management System implementation in hospital

Racha Dabliz, Jonathan Penm, Angus Ritchie, Simon Poon

Background: A widely used perceived workload measuring tool is the NASA Task Load Index (NASA-TLX). It uses six dimensions as measures of workload: mental demand (MD), physical demand (PD), temporal demand (PD), performance (PE), effort (EF), and frustration (FR).

Aims: We challenge this hypothesis by excluding PE and using the remaining five constructs as a measure of workload.

Methods: The NASA-TLX was distributed staff in the ICU (n=40) and Oncology Unit (n=27). CFA was conducted on the proposed models. The original six constructs were first tested against load (model 1). Model 2 replicated a model hypothesized by Tubbs-Cooley et al, and modelled MD, PD, TD and EF as measures of load. We challenged this hypothesis integrating FR, and tested a model that MD, PD, TD, EF and FR as indicators of load. Validated model fit tests used were RMSEA <0.05, CFI and TLI >0.95, and a SRMR < 0.08 and CR and AVE.

Results: The CFA results for model 1 illustrates that all six items didn’t fit the model well (CFI =0.953, TLI=0.922, RMSEA =0.103, SRMR=0.0652, CR= 0.44, AVE =0.44). Model 2 demonstrated a better fit, (CFI =.1.0, TLI=1.062, RMSEA =0, SRMR=0.074. CR= 0.87, AVE= 0.55) than model 1. Model 3 which introduced FR to the 4 items had a similar fit to model 2 (CFI =.1.0, TLI=1.389, RMSEA =0, SRMR=0.027, CR/AVE=0.53).

Conclusions: This study also showed that the measurement of subjective mental workload via the NASA-TLX tool needs to consider each of its dimensions rather than a global score.