--- title: "Why Existing In-House Data Rarely Gets IVD Software Through IVDR" description: "Discover why in-house data often falls short of IVDR clinical performance requirements for IVD software and how to choose the right study design." url: https://qbdgroup.com/en/blog/ivd-software-in-house-data-clinical-performance type: "Blog post" language: en published: 2026-10-14 author: "Pieter Smits & Pieter Bogaert" category: "Software Solutions & Services" publisher: "QbD Group" citation: "QbD Group, \"Why Existing In-House Data Rarely Gets IVD Software Through IVDR\", https://qbdgroup.com/en/blog/ivd-software-in-house-data-clinical-performance" --- # Why Existing In-House Data Rarely Gets IVD Software Through IVDR > Discover why in-house data often falls short of IVDR clinical performance requirements for IVD software and how to choose the right study design. "We already have so much data. Do we really need an expensive external clinical study?" It is the question we hear most from IVD software manufacturers. For novel devices, the honest answer is almost always **yes**. Existing in-house data can provide valuable evidence. But under the IVDR, it needs to demonstrate that the device achieves its intended purpose under the conditions in which it will actually be used. And that is where in-house data often falls short.

In this blog post:

## Clinical Performance Has a Specific Meaning Under IVDR Under the IVDR, clinical performance must demonstrate that your device, in its final design, achieves its intended purpose when used by the intended user in the intended use environment. Most in-house data fails at least one of those requirements.

It may have been generated:

These distinctions matter because **evidence generated during development does not automatically demonstrate clinical performance under the intended conditions of use**. Literature and routine diagnostic data are valid sources under Annex XIII, but for a genuinely novel algorithm, they rarely exist. The same goes for equivalent devices. So, for new-to-market software above Class A, **a clinical performance study is usually the linchpin of your evidence**. ## "Aid to Diagnosis" Does Not Lower the Evidence Bar The place of the device in the overall clinical workflow is an important factor when determining clinical evidence requirements and assessing the acceptability of residual risks. But describing software as an "aid to diagnosis" is not a free pass. It does not remove the need to meet applicable evidence requirements or justify below-par performance. **The evidence must still support the device's intended purpose and its role in the clinical workflow.** ## Usability Validation Is Part of Clinical Performance Evidence For standalone IVD software, the device interface includes more than the graphical user interface (GUI).

It also encompasses:

Your summative usability study, conducted in accordance with IEC 62366-1, contributes to clinical performance evidence. This is an important consideration for software manufacturers. **A usability failure is a performance failure.** Demonstrating that an algorithm produces the correct output is not enough if the intended users cannot use the software or interpret its results correctly. ## Does an IVD Clinical Performance Study Have to Take Place in the EU? No. A clinical performance study does not necessarily need to be conducted within the European Union. However, the specimens used must reflect the intended population. Manufacturers must also justify that the study results and conclusions are applicable to EU patients. **The key consideration is therefore not simply where the study takes place, but whether its evidence supports the intended use population.** ## Choose Your Clinical Study Design Based on the Device There is no single clinical performance study design that works for every IVD software product.

The appropriate approach depends on device-specific factors, including:

Depending on those factors, the study may be:

The important point is to select the study design based on the characteristics and evidence needs of the device, **not on theoretical grounds**. Plan the study accordingly. ## Build the Right Clinical Performance Evidence for Your IVD Software Existing in-house data can be valuable, but it does not automatically provide the clinical performance evidence required under the IVDR. For novel IVD software, demonstrating performance under the intended conditions of use often requires a dedicated clinical performance study. QbD Group supports IVD software manufacturers in defining clinical evidence requirements, selecting appropriate performance study designs and planning studies aligned with their device's intended purpose and regulatory pathway. --- Source: https://qbdgroup.com/en/blog/ivd-software-in-house-data-clinical-performance — © QbD Group. Quote freely with attribution and a link back.