Today, medical decisions may often rely on information provided as output from a software program meaning the software will be considered an IVD medical device underlying IVDR requirements. These IVD software products utilize algorithms processing input data (e.g. patient data, IVD test results, images) to produce output data. These outputs are intended to provide information on the patient for example the health status fulfilling a medical purpose (e.g. diagnosis, treatment decision or risk assessment). The results will affect the clinical decision-making and finally patient care and clinical outcome.
Thus, manufacturers of IVD software products are expected to fulfill the IVDR requirements incl. a detailed performance evaluation looking specifically on the software’s scientific validity, analytical performance and clinical performance. The scientific validity is the association between the software results/conclusions as based on the input parameters (e.g. clinical analytes, physiological data) and the used algorithm, with the target clinical condition (medical purpose). The analytical performance is provided by analytical validation, which confirms that the IVD software correctly processes input data to generate accurate, reliable and precise output data. Lastly, the clinical performance is the ability of the software to yield results that correlate with the intended clinical condition or physiological state in accordance with target population and intended user.
Establishing those evidence documents for the performance evaluation of an IVD software might be new and challenging. IVD software developers/manufacturers need to establish structures and processes to fulfil these requirements.
knoell experts can help you to identify and close your current gaps to those new requirements and support you in developing a quality management system by providing skilled expertise and process know-how