Quantitative Exploration of Nurses’ Task-Technology Fit within the Electronic Health Record Using Regression and Data Mining: An Ehr Workflow Analysis

Abstract

BACKGROUND: As we continue to progress with the use of Electronic Health Records (EHR), there is a need to improve nurses’ user experience (UX). An understanding of nurses’ task-technology fit (TTF) with EHR performance and an examination of nurses’ navigational behaviors within the EHR is warranted. PURPOSE: To quantify nurses’ TTF and describe their navigational patterns in the EHR to provide pragmatic solutions to improving nurses’ UX with the EHR. METHODS: Using a cross-sectional quantitative design, the first phase of the study examined the relationship between individual, task, and technology characteristics to subjective workload, EHR efficiency, and composite performance. The second phase of the study used computational ethnography using EHR audit logs as an in-situ data source. Sequential Pattern Mining (SPM) and Markov chain Analysis (MCA) were used to examine the audit logs. RESULTS: In a sample size of 95 nurses, 50% were white, 84% were females, 83% considered themselves as EHR proficient, and 34% worked in a Medical/Surgical unit. The first phase of the study revealed age and nursing experience were negatively correlated to ease of use/training, workload, and efficiency. Regression analysis also showed that the relationship with informatics staff was a strong predictor of nurses’ EHR workload. The second phase of the study was from a random sample size of 20 nurses from the Phase I participant pool. “Documentation” was the most frequent EHR event that was used by nurses accounting for 28% of their EHR time. SPM revealed that “query clinician” – “documentation” was the most frequent sequential pattern. In reviewing the transition probability matrix, nurses were most likely to navigate launching an application after logging in. Two different models were presented, showing the primary and secondary navigational pathways that nurses would traverse when using the EHR. CONCLUSION: The study identified the different TTF characteristics that may impact nurses’ performance within the EHR. Healthcare organizations need to deliver a good TTF to attain an effortless UX. Unmasking time-based navigation behavior of nurses using computational ethnography can assist in the redesign of EHR screens to the right nurse, at the right time, and in the right sequence.