A literature review is a formal collection of scientific studies published in peer-reviewed journals or conference proceedings that are relevant to the safety evaluation of a medical device product.

Literature reviews are essential for manufacturers to ensure that the IVDR has been constructed according to current regulatory requirements and that it is based on the latest scientific knowledge. In addition, they allow manufacturers to understand better how other companies are approaching similar problems, which can be helpful when looking for new solutions or thinking about how to expand into new markets.
The usefulness of literature reviews for clinical evaluation report (CER) program management is evident. Literature reviews allow us to identify gaps in the technical knowledge about a medical device so that we can develop new or improved products, therapies, or procedures. In addition, literature reviews can support evidence-based decision-making by comparing the safety and efficacy of competing for medical device, assessing the risk-benefit profile of new products, and identifying gaps in knowledge that would benefit from additional research (e.g., clinical trials).
Manufacturers usually depend on a thorough literature review to highlight areas where there are conflicting results or no clear consensus among published studies; these findings should be taken into consideration during decision-making processes (e.g., regulatory approval). It is also a way of reducing costs by avoiding duplicating research that has already been done by another company or organization.

Needles in haystacks


literature review , systematic review process , systematic review , data extraction , automating data extraction , clinical trials , evidence based medicine, systematic reviews ,support systematic reviews


More than 3 million scientific articles are published in English every year, which is growing by almost 10% annually.

This information explosion makes it increasingly difficult for researchers to keep up with the latest developments in their fields. In addition to the apparent challenge of keeping up with new content, there is another problem that many researchers face: understanding the context, methods, and goal of the research being presented. Those can often only be understood by reading the papers in detail.
Hence, the regulatory team members are required to read, organize, and report on all of these articles. However, the high addition of papers to the scientific community has created an information overload. Given this rapid growth in the number of research articles, it is hardly possible to read or even skim through all of them to find relevant information for clinical evaluation reports (CER); or PER for in vitro diagnostic regulation devices.

Performance evaluation process and the literature review process

The IVDR’s performance evaluation report (PER) program is essential to Medical Device Regulation (EU MDR). It allows for assessing and predicting patient experience, outcomes, and satisfaction. In addition, the review process builds the foundation for the later tiers of the program.
Most current PER programs, however, have their own set of problems that make it challenging to create effectively, update and comprehensively summarise and present data. Usually, they are conducted by a team of researchers who manually search for and read relevant articles, use software programs to extract data from those articles, and summarize their findings in an organized manner.
The process is time-consuming, prone to human error and bias, and does not scale well for large amounts of data. Moreover, the extensive amount of research aside, organizing the search outcome in a relevant manner and updating them periodically, especially for multi-product manufacturing teams, is also a challenge.
As of now, data collected from research evidence are primarily kept in spreadsheets or combined files. However, a few critical shortcomings are noted in this method.
First, the reports are often not centrally located, and if they are, then they require management by a dedicated team, forcing the company to expend resources and time. The data is also extracted and stored in inconsistent methods, so there is no easy way for senior management to access all the information in one summarised space.
There is some use of software, especially to define the parameters, add term refinement and determine the types and data extraction methods. However, non-specific software is ill-equipped to handle the level of excellence required by the MDR and notified body. Not to mention, these systems usually have limited access to customizations.


performing systematic literature reviews, software engineering , living systematic reviews ,systematic review automation technologies, medical device , literature review process , international conference, machine learning methods, performance evaluation process


The use of artificial intelligence is not novel. AI allows for an excellent collaboration between human and machine effort, the very best of which is yet to come. Nevertheless, the currently available software holds the capacity to reduce workload by several times for manufacturers preparing literature reviews and performance evaluation reports.

The upside of automating the literature review process is more than a few. First, it can be argued that one has more control over the data screening process and data extraction in an automated process than in a manual process.

For example

  • a manufacturer can opt for automatic searching for relevant literature based on user-defined criteria;
  • automate the extraction of relevant data from the literature;
  • get auto summarization of results into an organized document or report; Distribute information to stakeholders through a web portal or mobile app in real-time;
  • Produce fully customizable reports based on their preferences and data needs.

Several Automated programs with integrated artificial intelligence and machine learning allow users to create labels, filters for the data extraction process, inclusion criteria & research questions customization, and indexing to balance precision and recall. Some come with summarisation and reporting tools highlighting hidden connections in the textual data. Automated bias assessment and study identification ensure the automated approach does not need much human intervention beyond the search strategy.
Interactive machine learning systems are more straightforward than one would assume. For example, some software offers data visualization, meta-analysis, evidence synthesis, and text mining.
The automated systematic review process has an edge compared to the manual process.

To sum up

The focus of PER, as mentioned earlier, is on evidence based medicine, so it is vital to retain this in mind. In the case of writing PER forms manually, the process can be time-consuming, error-prone, and may result in inconsistent evaluations. Since PERs are often used to make important decisions about resource allocation and funding, it is crucial that they are accurate and consistent to ensure successful resource allocation and funding decisions. Cost-effectiveness must be considered when proposing a solution. When used correctly, systematic review automation tools can effectively streamline the IVDR PER writing process.

Want More EU MDR and Regulatory Insights?

We send weekly emails with the latest regulatory developments, templates, and strategies straight to QA/RA Professionals like you. Sign up below to get access today.