What is it?
The basic idea of a computerized adaptive test (CAT) is to maximize the information and precision by adapting the questionnaire to the individual patient. This is done by using the responses to the previously posed items to select the most informative next item following the general principle, that if a patient has reported few problems on the previous items, the next item will concern a more demanding task, while if severe problems have been reported the following item will concern a less demanding task. The administration of items proceeds until a predefined level of precision or a predefined number of items has been reached
Such dynamic assessment has several advantages from a measurement perspective: it increases precision and measurement range, reduces floor and ceiling effects, and makes it possible to reduce sample size requirements. The measurement system becomes more flexible, and results from studies that have asked different questions are still comparable. Patients benefit from more relevant items and more effective use of the information they provide.
To improve measurement of a number of core HRQOL domains the EORTC QLG has developed a CAT version of the EORTC QLQ-C30 questionnaire. It measures the same well-validated and well-known domains as the QLQ-C30. For each QLQ-C30 domain a so-called item bank, from which the CAT selects the items for each patient, has been developed. The new items have the same item and response format as the QLQ-C30 items. The new instrument encompasses a total of 262 items. As the QLQ-C30 items are included in the item banks, the CAT is backward-compatible with the QLQ-C30.
How does it work?
EORTC CAT can be used in two ways, either as described above, or as so-called short forms that can be integrated in traditional paper questionnaires while preserving many of the advantages of CAT (described further below).
Assessment using the EORTC CAT requires online access as the program taking care of all CAT related computations is placed on a central server. All EORTC CAT assessments use this ‘CAT-engine’ for the required computations, but the presentation of items and storing of responses may be handled by different electronic data collection systems. The EORTC has a system available for this but other solutions may be used as long as these can be set up to communicate with the CAT-engine (the feasibility and requirements will be clarified in each case).
The general design of a CAT assessment is specified in a CAT-setting. This setting specifies which domains to measure, the start items, the number of items to ask, etc. The CAT program will then ask items in accordance with this setting until the selected stop criteria are fulfilled. When an assessment is completed scores are immediately estimated and stored.
If online access is infeasible for all or some patients in a study the EORTC CAT item banks may be used to generate so-called short forms. These are lists of items selected from the item banks that can be used as traditional (paper) questionnaires. Results from such short forms are fully compatible with the CAT measures and preserve many of the advantages of CAT. Responses from short forms are scored using a specialized EORTC program.
How is it used?
The conditions for using the EORTC CAT Core or EORTC CAT Core short-forms currently follow those for the EORTC QLQ-C30, although these conditions may be elaborated in the future.
The conditions for use of the EORTC CAT Core are:
- Completion of a User Agreement * EORTC CAT is used along with the QLQ-C30
- The applied CAT setting and/or short forms should be agreed with the EORTC QLG
- EORTC CAT should be scored according to the official algorithms (as provided by the EORTC QLG)
- The specific design and content of CAT-settings and short forms for a study will be agreed and set up in close collaboration between study and EORTC CAT representatives.
Questions about the EORTC CAT Core and use of the instrument can be sent here