Developing nurse-sensitive outcomes in acute inpatient mental health settings—A systematic review

Background: While nurse-sensitive outcomes (NSOs) are well established in numerous health settings, to date there is no indicator suite of NSOs for inpatient mental health settings. Aim: To assess the relationship between nursing variables and patient outcomes in acute inpatient mental health settings to determine which outcomes can be used as indicators of the quality of nursing care. Methods: Databases accessed were CINAHL, MEDLINE, PsycINFO and EMBASE, last searched in May 2022. The review followed the 2020 PRISMA checklist for systematic reviews. Papers published between 1995 and 2022, conducted in acute mental health care units were included. The quality of the studies was assessed using the Effective Public Health Practice Project Quality Assessment Tool. A meta-analysis was not possible because of the large number of variables and measurement inconsistencies. Results: A total of 57 studies were reviewed. Studies were categorised according to whether they found a significant or non-significant relationship between nurse variables and patient outcomes. Seven outcomes—aggression, seclusion, restraint, absconding, pro-re-nata medications, special observations and self-harm—were identified. For each outcome, there were significant findings for several nurse variables indicating that all included outcomes could be used as NSOs. However, evidence for aggression, seclusion and restraint use as suitable NSOs was more robust than the evidence for self-harm, absconding, pro-re-nata medications and special observations. Conclusion: All the seven outcomes can all be used to develop an NSO indicator suite in mental health inpatient settings. More work is needed to establish high-quality studies to clearly demonstrate the relationship between these outcome measures and changes in nurse variables such as nurse staffing, skill mix, work environment, nurse education and nurse experience. Patient and Public Contribution: Patient or public contribution was not possible because of the type of the variables being explored