Factors Influencing Online Nursing Activities

The primary goal of this study was to better understand how nurses interact with computers in the performance of nursing activities, more specifically how nurses create, modify, and discontinue computerized nursing orders for patient care. The main focus of the study was to a) understand how nurses currently used a text-based system to create, modify, and discontinue nursing orders, b) determine the differences in nurses' speeds, accuracy, and subjective satisfaction with text-based (CI-ICS) versus a redesigned graphical user interface in nursing orders management and c) examine the relationship of nurses' demographic and cognitive information processing characteristics during on-line nursing intervention orders management using the two different computer interfaces. The study was conducted at a large military medical center in the Pacific for purpose a) and at a large military medical center in the Mid Atlantic for purposes b) and c). For study purpose a) content analysis was used to analyze 14 nurses' comments about CHCS nursing orders. For study purpose b) a within subjects design allowed for a two-factor repeated measures analysis to be used, and for c) multiple regression was used to analyze the data in a descriptive correlational design.

A random, stratified (by gender) sampling technique identified potential nurses for the study purposes b) and c). A total of 98 clinical nurses participated in the main study. The majority of nurses were medical-surgical and intensive care nurses; however, the representation included psychiatric, operating room, and nursing anesthesia personnel. The modal educational level was a BSN and the average age of the sample was 33.9 years.

CHCS nursing orders were redesigned into a GUI interface and both interfaces were tested for nurses' interaction speeds, accuracy, and satisfaction. The GUI was significantly faster to learn and nearly twice as fast to use for overall speeds as well as after substantial practice for creating, modifying and discontinuing nursing orders. In addition to interaction speeds being faster, error rates were significantly lower with the GUI interface than CHCS. Even with practice, CHCS error rates were still high, meaning nurses lose productivity and commit substantially more errors more frequently with CHCS. Third, nurses rated the GUI interface significantly higher than CHCS for interface satisfaction scores. Users felt strongly about CHCS, as one participant wrote, "I'd rather stick needles in my eyes than use CHCS all day."

Using stepwise regression, the set of nurse characteristics explained 37.7% of GUI interaction speed and 31.2% of CHCS speeds. For overall speeds, the four predictors were significant. The largest predictor for GUI was previous GUI experience (I 8.6%). Age predicted an additional 11.7%, spatial visualization, 4.6%, and gender 2.8%. For overall CHCS interaction speeds, four variables were significant using stepwise regression. Spatial memory explained the largest amount of variance at 16.5%; previous CHCS experience explained an additional 7.5%. Then gender and spatial visualization added 3.8% and 3.4% respectively.

After practice, the set of predictor variables explained 27.7% of GUI interaction speeds and 24.3% of CHCS speeds. With GUI, stepwise regression showed that GUI experience was still the largest predictor at 18.2%; spatial memory added 5.3% and gender explained 2.7%. For CHCS practiced speeds, spatial visualization explained the largest amount of variance at 10.9% age explained an additional 6.6%, CHCS experience explained 3.5% more and gender explained 3.3%. The set of characteristics predicted little variance for error rates. Only age explained a small amount of GUI error variance, 9.4%. The predictors were not significant for CHCS error rate or subjective screen satisfaction.

The qualitative portion of the study was completed at a site that extensively uses CHCS. However, this study found severe underutilization of CHCS for nursing orders management. The under-utilization of CHCS is a significant problem for military nursing. CHCS is an expensive resource; its underutilization means millions of dollars have not been well spent, at least for nursing orders management tasks. The redesign of orders management into a GUI would allow for faster, more accurate orders completion and potentially help alleviate the underutilization of a large system.

The results of this study have direct and immediate application to military nursing. First, the results are a clear indicator that a GUI interface saves nursing time and prevents errors compared to CHCS for nursing orders management. In fact, using a moderate number of orders generated over 24 hours, an equivalent of 1-1.5 hours per day per nurse is saved. Likewise, if the military continues with CHCS, leaders must expect slower on-line speeds and higher error rates. Because CHCS is now deployed in the field setting, these same poor rates are now available in theater.

The high error rate with CL-ICS even after substantial practice is a grave concern. After 60 trials, nurses still had significantly more errors with CHCS than a GUI. More impressive, significantly more errors are committed with CHCS all through the teaming process at three times the rate of errors in GUI. Error rates with the GUI start and stay low compared with CHCS. The fact that nurses use CHCS worldwide makes it a seeming imperative to redesign and deploy a more modem and error-free system.

The fastest CHCS use is correlated to higher scores on cognitive variables, meaning fast use is related to higher use of cognitive resources. GUI, on the other hand, is better predicted with previous GUI experience. This has implications for military nurses in that the more nurses use GUls, the faster they become. Improving nurses' cognitive make-up is, of course, less amenable to change.

This study provides the first objective data about differences between text-based and GUI systems. Despite the widespread use of GUls, no empirical studies were located that documented the differences between the two screen interfaces.

The GUI program constructed for this study is available to CHCS 11 designers as a prototype, potentially lessening design costs for orders management in CHCS 11. More important, the study has application for priority-setting among clinical system initiatives. This objective data provides leaders with the ammunition needed to discriminate among the many competing initiatives in Defense health computing. Orders management should be redesigned and deployed in the near future to positively impact provider productivity and accuracy rates.

Final Report is available on NTRL: https://ntrl.ntis.gov/NTRL/dashboard/searchResults/titleDetail/PB2003102...