Predicting Research Utilization Among Registered Nurses in Alberta Hospitals Using Multilevel Analysis

Carole A. Estabrooks, PhD, RN, Faculty of Nursing, University of Alberta, Edmonton, AB, Canada, William Kai Midodzi, BSc, MSc, MSc, Knowledge Utilization Studies in Practice Unit, Faculty of Nursing, University of Alberta, Edmonton, AB, Canada, Rejean Landry, PhD, Laval University, Quebec, QC, Canada, Karen Golden-Biddle, PhD, Alberta School of Business, University of Alberta, Edmonton, AB, Canada, and Harley Dickinson, PhD, University of Saskatchewan, Saskatoon, SK, Canada.

Design: Data from a 1998 International Hospital Outcomes Survey conducted among Alberta registered nurses was used. Though rich in organizational variables, it lacked research utilization measures. To address this, we derived the dependent variable from a 1996 dataset of 600 Alberta staff nurses, and mapped this onto the 1998 dataset. This mapping was done using a regression model with statistically significant matching predictors between datasets.

Method: A hierarchical model, with nurses nested within nursing units within hospitals was used to partition and explained the variance in research utilization at three levels of individual nurses, nursing units, and hospitals. We used the model to investigate cross-level interrelationships, and quantify the amount of the variance in research utilization explained at the individual and organizational levels. We further obtained robust estimate of parameter estimate using Markov Chain Monte Carlo Simulation (MCMC).

Findings: Research utilization (RU) varies significantly at all three levels: individual nurse level [variance= 0.94 (91.85% of the total variance in RU), p<0.001]; nursing unit level [0.037 (3.62%), p=0.0005]; and at the hospital level [0.047(4.53), p=0.0063]. Significant factors found to predict research utilization among nurse were: nurse education, control over practice, freedom to make decisions, emotional exhaustion, and good nursing leadership on unit.

Conclusions: Variation in research utilization was mainly attributed to differences in individual characteristics of nurses, with organizational factors contributing less. Of the latter, however, having good leadership on the unit was the most significant determinant of research utilization Implications: Although organizational determinants explain less of the variance in our model, they are still statistically significant when analyzed alone, and may constitute the threshold needed before individual determinants can exert their considerable influence

Learning Objective #1: Identify factors that predict research utilization independently among nurse clinicians
Learning Objective #2: Identify what the most significant determinant of research utilization is