The fundamental requirements of medical radiation-protection legislation concern:
• the need to evaluate patient doses;
• establishing and using diagnostic reference levels (DRLs);
• paying particular attention to quality assurance (QA).
Healthcare facilities that undertake diagnostic X-ray examinations are required to establish DRLs for selected procedures and to carry out appropriate investigations and corrective action whenever they are consistently exceeded. A one-size-fits-all approach has been adopted for adult DRLs, which are defined for standard-sized patients of roughly 70 kg in weight.
Reasons why DRLs may be consistently exceeded include poor performance of one or more of the equipment components in the imaging chain, or the way in which the equipment is used. QA programmes are intended to reduce the probability of exceeding DRLs. However, the dose delivered during an X-ray examination is in itself a powerful routine QA parameter, since it is dependent upon the performance of both the operator and the X-ray equipment.
If patient dose assessments are to be effective as a QA measure, they need to be fully integrated into routine clinical practice and form part of a more general QA information-gathering process. IT-driven radiation-protection mechanisms could help to achieve this aim in a highly cost-effective manner.
To this end, a project was started in 1992 in Liverpool, UK, to develop IT-driven QA software tools centred on a modular Quality Assurance Dose Data System (QADDS). The software included a patient entrance surface air kerma (ESAK) calculation module that employs calibrated X-ray tube exposure factors, as well as specific patient-related examination details. Data from other dosimetric methods could also be entered. The tool also included QA modules for logging all necessary equipment performance data arising from departmental QA programmes.
Patient dose assessments employing calibrated X-ray tube outputs are extremely cost-effective.1 Calibration of the tube and generator is already a standard procedure for acceptance, status and periodic performance evaluation. Thus, use of the calibration data for patient dose measurements is effectively "value added": the cost per patient dose measurement decreases inversely with the number of measurements.
Early implementation
The QADDS package was installed in hospitals throughout the UK's Mersey region on standalone PCs for evaluation and testing throughout the 1990s, as recorded in a number of publications.2-4 Based upon experience gained, the QADDS V1 package was redeveloped in 2003, with the help of funds provided by the UK government's Department of Trade and Industry as part of its SMART investment programme.
The enhanced tool, developed to ISO TickIT software standards, was designed to be web compatible so that users could easily access it anywhere in their department. This approach also meant that any modifications, upgrades or servicing could be universally implemented with minimum disruption. All hospitals within the Mersey region have now been transferred to the new package.
In the V2 package, the ESAK module employs X-ray tube calibration data that can be updated automatically following an equipment survey. Manual data entry of patient examination details, especially when many records were involved, was found to be time-consuming and prone to error with the V1 package. To overcome this problem, direct access to radiology information system (RIS) data has been established for V2. This means that staff are only required to enter details once as part of an examination record (a legal requirement) and the patient dose assessment represents a "value added" outcome from this effort.
To verify the quality of the RIS data entries, a data audit software package has been developed to verify that any data entry is "reasonable" and useable. This RIS audit procedure corresponds to an important component of any clinical audit programme since it assesses the quality of the X-ray examination records. Direct access to radiological examination details recorded on the DICOM header of an image set for use in a dose estimate is equally possible.
The web-based approach means that all QA and patient dose data can be stored centrally.1 This supports the assessment of patient dose values against local and regional performance criteria and the derivation of appropriate local DRLs. Because QA data are also stored centrally, any variations in equipment performance can coincidently be assessed in terms of possible effects on patient doses and image quality.
What's more, because such large data sets (many tens of thousands of records) are now possible, improved statistical analysis of the data is routinely employed. This means that statistically significant differences in doses delivered to different patient populations can be highlighted and used as indices of examination quality, or in more detailed optimization and/or risk–benefit studies.5
Examination details downloaded on an ongoing basis from the RIS also means that patient dose measurements can now be employed as part of a routine QA programme. Running mean values of doses and exposure factors employed for particular examinations, as well as any statistical variations, provides powerful process-control measures at the individual X-ray room, equipment type, local hospital, regional or national level.
Routine knowledge of patient weight, sex and age can be used to define clinically related QA indices with minimal cost overhead.5 For example, the differences in mean doses for male and female patients provide valuable information on the performance of automatic exposure control (AEC) devices. However, studies have highlighted the need to establish a universal "gold standard" data set that can form the basis of X-ray examination records. Such records can then permit more detailed assessments of radiological practice as part of optimization strategies.
At present, routine RIS-based audit programmes are being established at a number of hospital sites. Once these are fully operational, IT-driven QA methods will be integrated into clinical practice as part of an automated cost-effective QA programme. This will help to standardize practices and save valuable radiographer time - as well as providing useful resource-management information.
Extended reach
One of the great strengths of a web-based approach is its capability to bring together standard data sets from many different locations worldwide. For example, a project underway at the University of Liverpool, in conjunction with PhD student Eric Ofori, involves QA and patient dose data collected in Ghana. Results can be collected to a standard format from anywhere in the world where internet access is available.
In addition, IT-based systems can help provide scientific support and encouragement to people working in environments with limited resources, helping them to become integrated into a much larger "virtual" scientific community. For example, cost-effective calibration of a tube and generator in, say, Africa, can be undertaken by means of thermo-luminescent dosimeters (TLDs) provided from a remote centralized laboratory and then employed to assess radiological performance for a large number of exams.
The application of IT to patient protection in diagnostic radiology has opened up the potential for new, improved and cost-effective scientific support mechanisms. Information generated can underpin the medical physics community's development of a more quantitative risk–benefit framework for diagnostic radiology, as well as optimized feedback loops or expert systems support for operators.