Stability of nursing home quality indicators over time

Nursing home quality indicators (QIs) provide a way to support quality assurance and improvement activities and to help ensure that cost savings are based on increased efficiency and not on decreased quality of care.

QIs values are expected to change over time. However, to be good indicators of quality, they should be reasonably stable over "short" periods. This paper discusses theoretical and measurement issues affecting stability and examines the stability of QIs over each of two 3-month periods and one 6-month period.

The study sample included 512 nursing facilities from two states, Kansas and South Dakota. QIs were measured for the first 3 quarters of 1996.

Facility level QIs were constructed using three different metrics that each provide a unique perspective of facility performance as follows: the proportion of residents in the facility with the QI condition; the facility's percentile rank in its state; and a variable indicating whether the facility's rank exceeded the 90th percentile in its state. QI stability was assessed using Pearson correlation coefficients, Spearman rank order correlation coefficients, and Cohen's Kappa, as appropriate for the metric.

Results indicated high levels of stability for most QIs, with lower levels of stability found to be in keeping with theoretical and measurement considerations.

QIs are reasonably stable over short periods of time. Quality improvement efforts may best be focused on facilities that are consistently poor performers over time, and those that show a large decrease in quality from one quarter to the next.