Abstract
Objective: In many hospitals, nurse-led "safety huddles" are used to relay patient safety information, although whether this effectively identifies patients at risk for harm has not been determined. New electronic risk assessment tools are designed to identify patients at risk for harm during hospitalization, based on specific markers in the electronic health record. This study sought to compare the results of both methods. The findings may help to enhance decision making at the level of care delivery.
Methods: A nonexperimental correlational study was conducted over a three-week period in 2015 in a large metropolitan acute care community hospital. Nurses on three units-a medical-surgical unit, a progressive care unit, and an orthopedic unit-constituted the convenience sample. Designated safety huddle leaders collected data using the daily census sheet to record the nurses' perceived risk of harm for each patient and the reason for risk concern. Separately, designated advanced practice nurses collected the electronic risk assessment tool's reports from the same units. Data were paired as they were entered into the database and analyzed to determine correlation. Perceptions of harm from the nurses, recorded as yes or no responses, were compared with the electronic tool's identification of high risk or moderate-to-low risk.
Results: In 746 data pairs, differences between the nurses' harm risk perceptions and the electronic tool's harm risk reports were statistically significant, supporting our prediction that there would be no correlation. The most significant difference was seen in instances when a nurse identified a patient as being at higher risk than the electronic tool did, often citing behavioral or psychosocial issues as the reason for concern.
Conclusions: Nurses perceived harm risk differently than the electronic tool did. In situations when the electronic tool cited risk and the nurse perceived no risk, the risks were currently being addressed in the plan of care. In situations when the nurse perceived higher risk than the electronic tool did, the nurse often cited behavioral or psychosocial issues (which frequently lacked defined data points in the electronic health record and thus were not available to the tool). Changes in data mining algorithms must incorporate and weight the impact of psychosocial and behavioral elements together with other risk factors in order to provide meaningful practice recommendations.