An Evaluation of a Sitter Reduction Intervention, Sitter Usage, Staffing Levels, and Patient Safety Outcomes

Purpose:
To examine the relationships among sitters (safety and behavioral health),restraints,falls,and staffing levels among 19 inpatient nursing units;to determine if staffing levels predict sitter type and usage,restraints,and falls; and to evaluate the effectiveness of a sitter reduction intervention.

Background/Significance:
Changes in the healthcare environment coupled with constrained resources have necessitated examination of how best to deliver patient care that requires patients to have more intense monitoring. Hospitals have utilized sitters (companions or constant observation) as an alternative to restraint application, to maintain a safe patient environment, to monitor patients for self-harm, and to reduce patient falls. Further research is needed to validate sitter usage in relation to patient safety.

Methods:
A sitter reduction intervention was implemented April 2011.The following tools were developed with input from shared governance councils and leadership:sitter justification assessment,sitter decision tree,family letter,sitter list,sitter justification form,and sitter evaluation form.Data were collected pre-implementation (September-November 2010)and three months(June-August 2011)after a sitter reduction intervention was implemented in regard to sitter type and usage, restraints,falls,and registered nurse(RN)and clinical care partner(CCP) staffing levels.Data were obtained from the supervisor safety report, NDNQI® reports,and KRONOS Analytics™.An electronic data collection log was used.

Results:
Pre-implementation, there were 18,917 sitter hours, $257,500 sitter cost, and safety sitters were related to staffing levels r=.50,p=.000.Behavioral health sitters were related to restraint usage r=.20,p=.001 and staffing levels r=.50,p=.000. Regression was conducted to determine if staffing levels predicted restraint/sitter usage and falls. The model predicted restraint usage R2=.39, R2adj =.37, F(4,1812)= 18.56, p= .001.Staffing levels accounted for 19.8% of the variance in restraint usage. Results indicated the model predicted sitter usage R2=.022, R2adj=.020, F(4,1812)= 10.12, p=.001. Staffing levels accounted for 14.8% of the variance in sitter usage. Post-intervention data are pending.

Conclusions and Implications for Practice:
Findings may provide leaders a better understanding of sitter usage in relation to patient safety and may strengthen the evidence base on how staffing levels and practice environments support patient safety. A sitter reduction intervention maybe cost-effective and maintain quality patient outcomes.