Evaluating the relative usefulness of old radiological equipment.

Abstract

When new therapies are evaluated in clinical studies, the results are expressed by variables that directly reflect patient outcome. Evaluation of the clinical usefulness of new or improved techniques in diagnostic radiology is not so straightforward. For specific diagnoses, the performance in terms of sensitivity, specificity, and predictive values can be evaluated by comparison with a gold standard. A different approach must be used when the aim is to evaluate or compare the general performance of radiological equipment. In this issue of Acta Radiologica, Hong et al. (1) present a study where the aim was to evaluate whether low-cost, used computed tomography (CT) equipment provided less diagnostic information than new, more costly scanners. An import ban on used medical equipment in South Korea was lifted in 1997, leading to substantial import of used CT machines. In line with several other countries, the fee-for-service per CT exam in South Korea is constant regardless of the age and quality of the scanner. The authors have had the impression that this might have caused an undue increase in the import of used CT scanners with limited diagnostic capabilities. The authors hypothesized that improved quality of the CT exams would reduce the need for repeated CT exams in the follow-up of the patient. This is an excellent example of the use of surrogate variables in the evaluation of performance in diagnostic radiology. The study indicates that when the CT scanner was acquired second hand (“used scanner”) the odds ratio for a retake CT was higher than for scans performed with machines that were purchased as new. Another challenge for the study was the need for stratification of several variables that may influence retake rate. The solution was to include a very large number of cases, approximately 250,000 CT scans. The study illustrates the power of large numbers in documenting statistically significant differences, even when the absolute differences can be small. In practice, such large numbers can only be obtained by extraction from databases. The down-side is that some interesting variables may be unavailable for analysis. In this study the clinical and radiological information on each scan was limited, and we do not know whether the protocols or radiation doses for CT scans were systematically different between facilities with new and used scanners. According to statistics from the Organization for Economic Co-operation and Development (OECD) (2), approximately one out of every 10 inhabitants in South Korea underwent a CT scan in the year 2010. The figures for several European countries are approximately similar, while the examination frequency in the United States is approximately 150% higher. Import of used radiological equipment is not common in the Nordic countries, although there is no import ban. This reflects the privileged economic situation in these countries. The situation is in striking contrast to the situation in many developing countries, where CT scans are practically unavailable for the majority of the population (3). The issues addressed in the study by Hong et al. can be of interest for countries with improving living conditions, when more advanced equipment such as CT machines will be installed. Several hundred CT machines are available for sale on the Internet; some of them are nearly 20 years old. Equipment selection is a complex balance between cost, image quality, radiation dose, throughput, availability of service and spare parts, et cetera. How these factors come into play in different countries is a very relevant research topic.