Clinical Evaluation of Software as a Medical Device according to MDCG 2020-1
Clinical evaluation refers to a process that must be performed as part of the quality management system in the life cycle of any medical device. In clinical evaluation, the clinical data of the medical device are assessed and analyzed to demonstrate its safety, efficacy and performance.
This also applies to Medical Device Software (MDSW), as it too can have a significant impact on the health of patients and the public. Thus, performance and safety must also be demonstrated for MDSW within its intended use and scope of application, and the risks associated with its use must be acceptable in light of the benefits to the patient.
MDSW differs from other medical devices in that it is part of a complex, highly interconnected, and interactive technical environment in which changes and modifications can be made more frequently and implemented more quickly or efficiently.
Because many MDSW do not directly impact or have physical contact with the patient, the clinical utility of MDSW is different than that of other medical devices: MDSWs are used to provide medical information that clinicians can use to make diagnoses or prescribe therapies. While the input data often comes from other medical devices or in vitro diagnostic devices, the medical decisions and thus the clinical outcome usually lie with the clinician.
Determining the clinical strategy
The following three types of MDSW differ in terms of their respective clinical strategies:
- software with its own medical purpose and clinical utility, e.g. MDSW that is able to identify tumors based on CT scan data. In this case, the clinical evaluation is prepared for the software only.
- software with its own medical intended purpose as part of another medical device, whose clinical benefit only arises in interaction with other medical devices and which only together achieve their intended purpose, e.g. MDSW for implantable cardiac rhythm monitors for long-term monitoring of electrical cardiac activity. Here, the clinical evaluation is performed jointly.
- software as an accessory that influences another medical device but does not have its own intended purpose and clinical benefit. e.g. a smartphone app for an insulin pump. The clinical evaluation is also performed jointly here
Three key components for deriving clinical evidence
According to MDCG 2020-1, there are three key components used to derive clinical evidence for MDSW:
1. Valid clinical association or scientific validity
Valid clinical association or scientific validity refers to the correlation between output data generated based on the input data and the algorithm of the MDSW and the clinical condition or clinical target parameters according to the intended purpose.
This correlation must be clinically accepted and scientifically justified. It is demonstrated by clinical performance data, e.g., from proof-of-concept studies or own clinical trials, taking into account the current accepted state of the art based on guidelines and/or the scientific literature. Published data such as those from the Summary of Safety and Clinical Performance (SSCP), as well as those from regulatory registries and databases may also be used. Clinical performance or compliance with standards may serve as additional input to the evidence.
If existing data are insufficient, additional clinical performance data must be generated, e.g., in the form of clinical trials/studies or through analysis of real-world data.
2. Technical / analytical performance
Technical/analytical performance validity means that the MDSW can accurately, reliably, and precisely produce the corresponding output data from the input data.
Technical/analytical performance is the basis of clinical performance and is demonstrated during verification and validation activities. It is part of the General Safety and Performance Requirements (GSPR). Verification and validation must be performed in the intended computing environment (e.g., hardware, storage size, network infrastructure, etc.) and in the intended context of use (real user interaction, required operating environments such as cloud operation or remote network operation, etc.).
3. Clinical Performance
Clinical performance is the ability of an MDSW to achieve clinically relevant results as intended.
To demonstrate clinical performance, the device or equivalent must have been tested in the intended patient population, user group, operational and usage environment, under the intended conditions of use. Validation of clinical performance must also demonstrate that the MDSW is user-friendly and capable of delivering the desired clinical outcomes envisioned in its intended use. Demonstration of clinical performance is based on product characteristics and may include preclinical testing, clinical trials, or a clinical performance study.
If the MDSW has multiple features, only one or a few of which provide clinical benefit, validation of clinical performance may be limited to these feature(s).
Special case modular software
For software that includes multiple modules, clinical performance can be validated for each module individually if the functionality of that module is independent of that of the other modules. This allows an independent risk assessment for a changed single module.
Both the quantity and quality of data are critical to the measure of clinical evidence and can be guided by the following questions:
- Does the data support the intended use, indications, clinical claims and contraindications?
- Have the clinical risks and analytical/clinical performance been examined?
- Have relevant features of the MDSW, such as data input and output, algorithms used, or how they are linked, been considered in generating the data to demonstrate performance?
- What is the degree of innovation or how long has the MDSW been on the market (this determines the extent of the required, scientific evidence)?
- Were the type and study design adequate with regard to the research objectives?
- Are the data used appropriate and up-to-date (state of the art)?
- Was the statistical approach appropriate to reach a valid conclusion?
- Were all ethical, legal, and regulatory considerations/requirements taken into account?
- Can a conflict of interest be excluded?
Conclusion and recommendation
The growing market for MDSW offers manufacturers new opportunities as well as challenges to meet regulatory requirements.
- Define the strategy for the preparation of the clinical evaluation at an early stage.
- Verify that all three key components of clinical evidence for your MDSW have been addressed in a timely manner.
- Determine whether the quantity and quality of clinical data for your MDSW under the MDR are sufficient to demonstrate the safety and performance and clinical utility of your product.