Turnover and Patient Satisfaction

¶ … Nursing Unit Turnover on Patient Outcomes in Hospitals While it might seem intuitive that higher rates of turnover within a unit are problematic, the study by Bae, Mark, & Fried (2010) attempts to show using a quantitative study the extent to which turnover has had a negative impact upon patient care. This study compared the levels of turnover at 268 units from 141 hospitals and assessed its relationship to both patient satisfaction and the nurse's own perceptions of cohesion, relational coordination, and learning.

The study is quite distinct in design from qualitative studies which often have an anecdotal approach and emphasize experience vs. data in evaluating policies. One problem, however, with the use of such a study which emphasizes numbers over narration is the difficulty of quantifying subjective variables like efficacy. In this instance, nurses' own perceptions of what policies were useful were assessed as well as patients' perceptions to contextualize both findings and enhance accuracy.

However, the question remains if nursing units with lower turnover were truly more cohesive or whether this was merely their perception. Patients may also not be the best individuals to use to evaluate whether they truly received superior care. Finally there is the inevitable question of correlation vs. causation. Poorer outcomes and poorer levels of morale amongst nurses may be associated with higher rates of turnover but the turnover is not necessarily the cause of the problem but might rather be associated with other variables.

However, some of these concerns are at least mitigated by the statistical tests used to validate the results. The specific test used to evaluate the results was that of an OLS model (Ordinary Least Squares). "Ordinary least-squares (OLS) regression is a generalized linear modelling technique that may be used to model a single response variable which has been recorded on at least an interval scale. The technique may be applied to single or multiple explanatory variables and also categorical explanatory variables that have been appropriately coded" (Hutcheson 2011).

The use of this technique suggested that the "clustering of nursing units within hospitals" had minimal impact upon results (Bae, Mark, & Fried 2010: 44). This meant that the overall characteristics of the hospitals was less of a significant factor than the overall level of turnover. The statistical tests designed to reduce concerns about correlation vs. causation regarding the factors which impacted the results thus seemed to be appropriate and germane to the research.

Inferential statistics attempt to identify overall trends in the macro population based upon an analysis of trends in the studied population. The inferential statistical findings suggested that nursing groups with a 1 point increase in workplace learning were 44% less likely to have medication errors; nursing units with lower levels of turnover were 24% less likely to have patient falls (Bae, Mark, & Fried 2010: 45-46). Based upon the studied population, workplace learning and low turnover had a strong positive association with improved patient outcomes in critical areas.

The implications of the study suggested that efforts to decrease.