Nursing workload measures and case-mix: an investigation of the reliability and validity of nursing workload measures

The general aim of this one-year project, funded by the Department of Health and in collaboration with CASPE, has been to investigate the relationship between case-mix and nursing workload measures. Specifically, this report focuses on the underlying issue of examining methodologies and instruments used for measuring workload and the assessment of the sensitivity of chosen measures. The relationship between diagnosis and the process of outcome of patient care is also described at the individual patient level and the overall relationship of case-mix and nursing workload is explored at ward level. The rationale for the choice of Nursing Workload Management Systems (NWMs) and the study design adopted to assess the NWMs chosen is described. Thus, data were collected on three wards over a 6-day period to generate workload estimates, actual nursing hours worked, case-mix (DRG) information, and outcome measurement, The intra-system differences described focus on the financial implications of the choice of parameters used to generate workload estimates. Inter-system differences are described in terms of comparison with actual hours worked and the financial implications of over- and under- staffing estimates. Although the correlations between workload estimates are high, there are variations which loosely reflect other known differences between the NWMs. Correlations between workload estimates of over-staffing, per capita hours paid, case-mix and outcome measurement were low, showing that there was little difference between three of the NWMs (FIP, EXCELCARE and SENS) and confirming that the Criteria for Care behaved differently. Analysis at an individual case level was conducted to see whether it was necessary to take account of diagnostic group. This suggested that there was a relationship between the main diagnostic group and the quality and outcome of patient care even after taking into account the strong association between diagnostic group and patient dependency levels, although this relationship disappeared after taking ward effect into account. The overall conclusion reached is that the NWMs reviewed produced inconsistent and unreliable estimates of nursing workload. These estimates also make insufficient allowance for the skill-mix manipulation required to deliver good quality care. Recommendations are made based on these findings.