Abstract
Introduction: A Level I trauma center routinely faced challenges with meeting data submission deadlines and frequently struggled with a backlog of cases that limited opportunities for concurrent performance improvement. To provide a validated algorithm through which registry workload could be evaluated, the study institution designed a scientific model that predicted the amount of time required for chart abstraction on a patient-by-patient basis.
Methods: As part of this quality improvement endeavor, registrars documented the amount of time required to complete each chart. A total of 600 patients' data were included by randomly selecting 150 patients from each of the 4 trauma registrars. Given that no previous study has examined the association of patient-related factors with chart abstraction time, study variables utilized to construct this predictive model were determined by the trauma program manager and the lead trauma registrar.
Results: Multiple linear regression demonstrated that inhospital mortality; transfer from a referring facility; hospital stay; ventilator days; and number of complications, specialty consults, injuries, blood products, and procedures were significant predictors of chart abstraction time. The equation for the regression line for the multivariate regression was as follows: Y = 38.95 + 31.28 x mortality + 15.33 x referring facility + 4.68 x complications+3.55 x hospital stay + 3.33 x consults + 2.83 x diagnoses + 2.00 x ventilator days + 1.78 x blood products + 1.09 x procedures.
Conclusions: The merit of this prediction model is that it is based on patient-related variables and predicts time on a patient-by-patient basis. This innovative tool can be utilized by other trauma centers to evaluate registry productivity and identify opportunities for improvement retrospectively.