The results of an analysis of the cognitive information-processing workload of patient monitoring technicians working in central monitoring units (CMU) of five hospitals in Oklahoma are presented. The analyses were conducted in June 2016 during a visit to the four of the hospitals. The analysis included conversations with the management team, nurses and monitoring technicians at all five hospitals. The objective of the study was to determine whether the information processing (IP) load exceeds the processing capacity of monitoring technicians, and whether any overload is great enough to necessitate consideration for IP load-reducing strategies and interventions. This determination was accomplished by identifying whether an IP load gap exists between the current load and an industry-accepted IP load limit. The healthcare literature appears to be sparse in this area, making difficult any determination that changes in error rates are a function of changes in IP load. However, it is assumed for this study that an increased IP load or length of exposure to the current load will result in an increase in error rates among technicians. Errors are defined (albeit forensically) in this study as events, such as missed alarms, late response to alarms, misinterpretation of alarms and/or EKG strips, missed alarm notifications of nursing staff, etc.
The alarm processing load for a technician is partially derived from a measured number of alarms collected at one of the hospitals between 12/23/2015–1/19/2016. During that time, there were 748,150 alarms noted. The load may be smaller at the other hospitals; however, fewer technicians on duty make the processing load similar across all five hospitals. There is a probability of 31% that an alarm will activate in any one operational second (0.31 alarms/second). A measure of the rest of the IP load was derived from the amount of additional data analyzed and/or managed by the technicians, such as EKG interpretations, hospital admissions and discharges, phone calls made to nursing staff, etc. The visual IP load per technician was calculated and determined to be approximately seven to eight bits/second (a "bit" is the amount of information necessary to decide between multiple alternatives (Freivalds, 2014)). This is compared to a recommended IP load limit of three to four bits/second for visual information processing (Miller, 1955). While the information load for each technician is high, it appears that the screen information density is reasonable, and that this may be able to be safely and effectively managed, but only in the short term (∼two to four hours). Two-to-four hour rotations should be considered (Chewning and Harrell, 1990; Goldstein and Gigerenzer, 2002; Marois and Ivanoff, 2005; and Nadav-Greenberg and Joslyn, 2009).
