Abstract:
|
In healthcare, assessing the clinical quality of care delivered by hospitals and physicians is essential to understanding how to improve care and outcomes for patients. However, without well-defined metrics of quality, it is very difficult to evaluate provider performance. In this paper, we discuss how we use a surrogate variable, risk-adjusted post-operative length of stay (RApoLOS), as an objective indication of complications or unexpected events resulting from the treatment and care received during the inpatient hospital stay. Bridging traditional regression modeling and the traditional moving range control chart in statistical process control, we are able identify potential outliers (those who stay longer than expected) for each hospital by accounting for each patient's risk profile. These outliers have proven to be highly correlated with the complications coded by hospitals, as well as with excessive costs of care. We conclude with a sampling of applications of this powerful surrogate variable to answer practical questions in healthcare.
|