Information ttechnology is keeping tabs on referring doctors' inappropriate study orders and could be useful in automatically tracking imaging follow-up to ensure radiologists' advice is heeded, according to presentations at an informatics session on Wednesday.

Massachusetts General Hospital researchers demonstrated a decision support program that identifies ordering patterns for CT and MR studies. This web-based system was implemented on the hospital intranet in the final quarter of 2004. When ordering a CT or MR, the referring physician notes the patient record, indication, clinical signs and symptoms, and results from previous tests. This information is mixed with patient demographic data from the hospital RIS. Using American College of Radiology appropriateness criteria, the program then generates a score of appropriateness indicating likelihood of a positive result. A color-code system depicts orders that may be inappropriate and unlikely to be positive as red. Those with a likely positive result and clinical utility are green. All data are stored in a data warehouse and can be sorted by indication, body region, and patient characteristics such as sex and age.

Among other findings, the information reveals:

• The highest red zone by age group in CT occurred in the 10 to 19-year-old range.
• There is a significant difference in appropriateness of ordering by body region. Spinal exams have the highest prevalence of inappropriate study orders, whereas chest exams had the lowest.
• Clinical indications with the highest prevalence of inappropriate study order patterns include dementia, syncope, chronic headache, and back pain.

The system allows the hospital to both identify inappropriate patterns and target them for corrective measures, said presenter Dr. Pragya Dang, a Mass General fellow.

"Over time, the red scores have decreased, and the number of exams with higher positive findings has increased. Prior to implementation of the system, we saw a steady volume rise in CT and MR. Since then, CT and MR exam volume has flattened," Dang said. In another study, researchers at the University of Maryland reported that referring physicians do not necessarily follow radiologists' advice for follow-up imaging studies of abnormal findings. This could potentially leave radiologists more vulnerable in malpractice cases, as higher awards are often given when there is a lapse in communication and follow-up of results. Of 56,083 imaging exams included in a 15-month retrospective study, 2.9% (1650) were abnormal. Referring physicians were contacted by telephone and notified of results and advised regarding imaging follow-up. Of the abnormal findings, 153 (almost 10%) had no evidence of imaging follow-up. The chain of communication has numerous pitfalls, said Dr. Amy Musk, a radiology resident who presented results. Teaching hospitals have frequent changes in providers of primary care treatment, and cases of missing patient data are widespread in hospitals generally, she said. It's possible that in some cases, patients decided against follow-up imaging studies, but this was not documented. An automated tracking system could help in ensuring that findings are received and acted upon by referring doctors, the researchers said.