Evaluation of Quality of Trauma Care in a Local Hospital Using a Customization of ASCOT.

Abstract

Evaluating processes of care and outcomes of injured patients are important if improvements in the quality of care delivered to injured patients are to be accomplished. We applied a customized ASCOT model developed from our database as a tool to criticize the quality of care in a local hospital. A total number of 4,096 trauma patients were used to develop new coefficients for customization of ASCOT. These patients were divided randomly into two equal groups. New coefficients for ASCOT were derived from the first set of patients. The newly developed model was validated in the second group of patients and the measures of discrimination (area under the ROC curve) and calibration (Hosmer-Lemeshow goodness of fit) were calculated. Then we used the customized model to calculate the W score in different subgroups of patients treated in a local hospital to evaluate the care offered to patients in each group. The customized ASCOT had a good discrimination (area under ROC curve = 0.9575) and calibration (Hosmer-Lemeshow goodness of fit p value = 0.7628) in the validation dataset. Using the customized model, we calculated W score in different subgroup of patients treated in a local hospital for a period of six months. The quality of care was worst for laparotomy (W = -13.31) and pelvic fracture (W = -5.56) and best for orthopedic (W = 1.76) operations. We believe that a customization of ASCOT model when used for evaluation of quality of care in a local hospital can be useful for detection of defects and improvement of the process of care delivered to the patients.