Quality assurance for abdominal CT: a rapid, computer-assisted technique.

OBJECTIVE

Maintaining high standards in a large CT imaging department with multiple scanners, a large technical and clerical staff, and a rotating staff of radiologists is an ongoing challenge. We undertook a project to design and implement a simple, rapidly performed computer-assisted system of quality assurance (QA) for use in abdominal CT. In our project, we also analyzed the results of that QA system.

MATERIALS AND METHODS

We graded 1810 abdominal CT studies done in a 50-week period, using a three-point scale to indicate the quality of the following five parameters of technical quality: IV contrast enhancement, oral contrast opacification, window settings and artifacts, conformity to radiologists' protocol, and completeness and accuracy of header and scout data. In addition, a parameter reflecting performance of the film library and clerical staff was similarly graded. To provide a measure of peer review for radiologists, any disagreements with prior CT study reports were recorded when comparison studies were reviewed in the process of CT interpretation. A commercially available spreadsheet and database software program was tailored to allow rapid, easily performed data entry and analysis. Tables and graphs showing performance of technologists and film library and clerical staff were generated. This customized program was made available on the radiology department computer network. Results generated by the program were further analyzed with linear regression models.

RESULTS

Our QA system was successfully integrated into the routine operation of the abdominal CT division. During the first 11.5 months of operation, the system reflected improvement in each of the technical parameters with a statistically significant improvement in the combined average technical score (from 1.15 to 1.68 on a scale of 0-2; p < .0001). The "Throughout Speed/Old Exams" parameter for performance of the film library and clerical staff, which was analyzed separately from the technical parameters, also improved significantly (from 1.3 to 1.8; p < .02). Improvements were statistically significant, even when we controlled for potential variations in quality among different CT scanners and variations among the radiologists who rated the quality of the examination. Thirty-eight disagreements with previous scan interpretations (5% of all scan comparisons) were recorded for evaluation at peer review conferences.

CONCLUSION

The ability to monitor performance continuously using a rapid, computer-assisted system has effected measurable improvement in our CT service. Technologist and film library and clerical staff performance improved for all parameters studied. Deficiencies were revealed and trends demonstrated. The QA system allowed us to identify disagreements in interpretation of CT examinations for subsequent peer review by radiologists. Our QA software program has been made available on the Internet as freeware to licensed Excel users via anonymous file transfer protocol at Internet Protocol 134.192.6.110.