Understanding how physician perceptions of job demand and process benefits evolve during CPOE implementation

ABSTRACT

We examine how physicians’ perceptions of two computerized provider order entry (CPOE) capabilities, standardisation of care protocols and documentation quality, are associated with their perceptions of turnaround time, medical error, and job demand at three phases of CPOE implementation: pre-go-live, initial use, and continued use. Through a longitudinal study at a large urban hospital, we find standardisation of care protocols is positively associated with turnaround time reduction in all phases but positively associated with job demand increase only in the initial use phase. Standardisation also has a positive association with medical error reduction in the initial use phase, but later this effect becomes fully mediated through turnaround time reduction in the continued use phase. Documentation quality has a positive association with medical error reduction in the initial use phase and this association strengthens in the continued use phase. Our findings provide insights to effectively manage physicians’ response to CPOE implementation.

KEYWORDS:

• Computerised provider order entry
• implementation
• standardisation of care protocols
• documentation quality
• process benefits
• medical error
• turnaround time
• job demand

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. We used a tool to determine the gender of a name (http://genderize.io/, last accessed on July 17 2022). The tool provides the probability of a name being male or female. We acknowledge the limitations of this approach: such tools tend to do better with names on which they are more extensively trained (e.g., western names), and determining gender of an individual is more complex than classifying the gender of a name.

2. “The Comparative Fit Index (CFI: Bentler, 1990) is a revised form of the NFI which takes into account sample size (Byrne, 1998) that performs well even when sample size is small” (Hooper et al., 2008). NFI underestimates fit for samples less than 200 (Bentler, 1990; Mulaik et al., 1989).

3. Constraining the correlation to 1 among construct pairs resulted in a significant deterioration in model fit, providing additional evidence of discriminant validity.

4. In our analysis, we controlled for the effect of ease of use that has been considered as a system quality dimension for the success of IS in many previous studies.