A prospective study was carried out to evaluate the effect of teleradiology on the diagnosis, treatment and prognosis of patients in primary care. A university hospital was sent 685 plain film examinations via an ISDN connection from a primary care centre, for a radiological report. The study was conducted in two phases: during phase 1 (446 cases) general practitioners (GPs) selected the examinations, and during phase 2 (239 cases) all consecutive examinations were transmitted. In phase 1, 40% of the examinations were of the chest and 24% were of the spine; the remaining 36% were mainly bone and sinus examinations. In phase 2, 28% of the examinations were of the chest and 19% were of the spine. The sensitivity and specificity of the GPs' interpretations (compared with the radiologists') were 0.85 and 0.62 respectively in the first phase, and 0.90 and 0.86 in the second. In at least one-third of all cases, teleradiology helped with the diagnosis, although completely new diagnoses were less common. An effect on treatment was noted in 15% of cases and on prognosis in 5%. However, an appropriate consultation level is required for these positive effects. Adequate accuracy and patient safety cannot be achieved if the examinations sent for radiological reporting are preselected by a GP.