contributing to decreased confidence in the validity and reliability of PCSs and the staff allocations that result from their use (Malloch and Conovaloff, 1999).
Workload estimates used in PCSs also are criticized as not taking into account other factors, such as the frequent interruptions encountered by nurses in performing certain tasks (Malloch and Conovaloff, 1999) or the need for multitasking by nurses—often in the performance of the invisible, cognitive work of nurses described in Chapter 3. For example, while a nurse may be changing intravenous tubing, he or she may also be observing the patient’s physical status and pain level and providing patient education (Malloch and Conovaloff, 1999). Workload estimates also are criticized for being derived from measurements of care that is delivered, which is often constrained by staffing limits and therefore is not an accurate predictor of the care that is actually needed (Jennings et al., 1989).
One study compared PCS predictions with the care actually delivered, as measured by the same classification tool administered retrospectively by nurses who had received intensive training on the use of the PCS tool and had scored high on interrater reliability. This study found significant differences in the average prospective and retrospective classification scores in two of the three nursing units in which the study was conducted. For all three units, the retrospective PSC scores were higher than the prospective scores. The times associated with these differences would result in staffing deficits of 0.24 FTEs, 0.72 FTEs, and 2.99 FTEs in the three units (Hlusko and Nichols, 1996). An earlier study comparing the application of four different PCSs for the same patient population found large statistically and clinically significant differences in hours of care needed by the patients according to those four tools (O’Brien-Pallas et al., 1992).
Such concerns point to the need to validate and evaluate PCSs during their actual implementation (DeGroot, 1994). However, although there are reports in the literature regarding the validity and reliability of a system during its initial implementation (efficacy), there is “a paucity of published research related to patient classification system validation after implementation” (effectiveness) (Hlusko and Nichols, 1996:40). Moreover, continual changes in personnel, work environments, tools and equipment, and technology in most workplaces result in corresponding changes in the time required to perform the work, necessitating revision of the standard times. As a rule of thumb, experts recommend that work measurement reviews and reevaluations be conducted annually and additionally on an ad hoc basis whenever major work redesigns are undertaken (Bayiz, 2003).
Multiple purposes create incentives for gaming. Although PCSs are used to predict staffing needs, they have other uses as well, such as to estimate long-term staffing requirements for budgeting purposes (Seago, 2002). These multiple purposes of PCSs provide incentives for “gaming” or ma-