Can Nursing Units With High Fall Rates Be Identified Using One Year of Data? Reliability of Fall Rates As a Function of the Number of Quarters on Which They Are Based.
Reliability-the extent to which multiple measurements of a target yield similar results-is critical in comparing healthcare provider quality. Hospital unit fall rates are widely tracked and used for benchmarking, but their reliability is not well-studied. Our twofold purpose was to estimate fall rate reliability, both in terms of signal (between-unit variability) relative to noise (within-unit variability) and in terms of the accuracy with which units can be classified as high-fall units; and to assess reliability as a function of the number of quarters of data used to compute fall rates. Using year 2013 data from 11,765 critical care, step-down, medical, surgical, medical-surgical, and rehabilitation units in 1,552 US hospitals, we identified high-fall-rate units, computed units' signal-noise reliability, and simulated data to assess accuracy of high-fall-rate unit classification as a function of quarters of data. When critical care units were excluded, median unit type signal-noise reliabilities for annual total and injurious fall rates, respectively, ranged from .74 to .82 and from .53 to .68. In simulation, seven quarters of data were sufficient to achieve top-decile misclassification rates at or below 10% for all unit types except critical care. Top-quartile misclassification rates were higher; even 12 quarters of data did not consistently yield top-quartile misclassification rates below 10%. In the absence of long-term data, and for units with low patient volume and unit types with very low fall rates, comparison with a unit's own historical data may be more helpful for quality monitoring than attempting to rank it among its peers. © 2016 Wiley Periodicals, Inc.
Research in nursing & health
Accidental Falls; Data Interpretation, Statistical; Hospital Units; Humans; Nursing Staff, Hospital; Patient Safety; Quality Indicators, Health Care; Reproducibility of Results; Time Factors
falls; healthcare quality; patient safety; quality measurement; reliability
Staggs, Vincent S. and Cramer, Emily, "Can Nursing Units With High Fall Rates Be Identified Using One Year of Data? Reliability of Fall Rates As a Function of the Number of Quarters on Which They Are Based." (2017). Manuscripts, Articles, Book Chapters and Other Papers. 476.