Paige Case Study: How a Customized QMS Helps a Medical Device Company Keep Their Agile Flow

Written by Rina Nir

Case studies | Confluence | Electronic QMS | JIRA | Testimonial

Analysis of pathology sample with Paige.ai's medical device, QMS

Paige.ai, a global leader in AI-based pathology diagnostic software, uses Jira and Confluence for their Quality Management System (QMS). RadBee Ltd. has helped this fast-growing startup create a QMS that supports their Software as a Medical Device products (SaMD) — including a custom extension for Confluence that alleviated a critical bottleneck from the release process and ensured compliance.

“I’m quite happy with where we are today because the system is highly customized to our needs — and we continue to tailor it as needed. We’ve minimized low-value items, such as maintenance and other repetitive manual work, to almost zero at this point. Now my team can focus on higher priority items, like continuous improvement projects or truly working with product engineering teams to move to the next level.”

Emre Gültürk

VP of Regulatory Affairs & Quality at Paige

Challenges

The company’s Software as a Medical Device (SaMD) products are updated frequently, and the manual documentation process — particularly creating Medical Device Records (MDR) index table— was painfully tedious work.
Risk of release delays due to the documentation not being ready on time.
Elevated risk of errors and nonconformities because of the manual process.
Risk of non-compliance and failed audits if each release doesn’t have a complete, accurate regulatory dossier.
Team members were bogged down by the documentation process with no time to take on high-value tasks.
No existing programmable Confluence extension or Atlassian Marketplace app to make the creation of index tables easier.

Solution

Radbee designed a tailor-made Confluence extension called “Powertags,” making creating MDR for new releases effortless and foolproof.
Powertags provide a robust MDR Index for the regulatory dossier, ensuring compliance over time.

Results

Easy release process that is aligned with the Paige release flow.
Passed audits by the FDA and other regulatory organizations with flying colors.
The Paige team has time to do high-value work, like improvement projects.
Increased the product and engineering teams’ speed by eliminating the need to duplicate Confluence pages.
Removed bottlenecks and paved the way for other optimized efficiencies with new features to further improve the quality system.
A QMS that fits a medical device Agile development process.
A close working relationship with Radbee means continuous improvements are fast and easy to implement, even during Paige’s off-hours.

bPaige.ai, AI based pathology analysis, QMS medical device

Challenges

Develop a QMS in Confluence that automates, streamlines, and speeds the medical device release process and alleviates the manual burden on the team.

When veteran quality and regulatory expert Emre Gültürk joined the Paige team in 2020, he inherited a project to overhaul the QMS for the innovative medical device startup. From his lengthy experience, Emre had a clear perspective on what makes a quality system tick.

“This is my eighth company, which means I’ve had the chance to build eight quality systems pretty much from scratch. I believe in a quality system that is actually usable by the users involved with the tool itself.”

This may sound basic, but keeping compliance apace with releases can be complicated for a fast-growing company like Paige. As a SaMD, compliance hinged on having a comprehensive MDR index with each product release. This regulatory dossier included ten categories of documents pertaining to a software version. Each category would then link to multiple Confluence pages, complete with page versioning details (i.e., effective date, official version number).

Powertag-MDR-Table, Confluence API, customize QMS medical device

When Emre joined the team, MDR index creation was manual. As a compliance expert, the first challenge Emre identified was the sheer frequency of new releases.

“Our software iterations are very rapid, and we have many iterations over a year. So, a year down the road, if I needed to go back and look for a software version at a point in time, how could I pull the exact, relevant content in given software documentation? That’s not easily doable with Confluence.”

Additionally, high opportunity costs were associated with having key development personnel mired in menial, labor-intensive tasks.

“I don’t need my team to spend time on maintaining a training system or moving electronic records from A to B. That’s not valuable work,” Emre says.

Still, the cost of developing a customized system was a consideration. There are no apps or plug-ins in the Atlassian marketplace to create these MDR index tables. Weighing the costs/benefits was part of Emre’s job.

“Agile medical device companies move at a fast pace and need a tailored QMS. Otherwise, they will be in the limbo of not focusing on the important work — they’re too busy maintaining and doing low-value, manual work to keep things flowing. So that’s why it’s important to invest upfront in customized systems.”

In Emre’s eyes, the bottom line was to get his team out from under work that wasn’t moving their main business forward quickly with an eye on both compliance and higher-value outputs.

Solution

Develop a custom Confluence extension that automates the MDR index creation process as much as possible to eliminate errors and ensure robust, traceable documentation that passes audits while supporting continuous improvement.

With these goals in mind, Emre and his team developed clear guidelines for Radbee to create a custom-built Confluence extension to fix the problem.

“We gave the specs to Radbee and stated that we needed robust, fully traceable index pages. When I open a given index page, it needs to pull the documentation for that specific version of the software.”

Leveraging the power of the Atlassian API, Radbee set to work on designing a solution called “Powertags.” Each page gets new tags: one tag to identify the product version and a second to associate it with the category of the document in the MDR table. This is the only step that the page author must do “manually.” These tags are linked to a specific version of that page and are maintained even when the page has newer versions.

Then, when a release is ready, a new MDR Index page is created using a blueprint template. The index table is completed automatically from all the Powertagged pages.

“RadBee delivered an impressive solution for us. Now our team simply generates an index every time we release software, and we establish those index directories in our medical devices records,” Emre says.

The automated system makes things not only faster but also more efficient and accurate.

“If the system didn’t exist, we would duplicate the software documentation for every single iteration because that’s how we would maintain the different versions in between them. Enabling something like this increases our product and engineering team speed, as we don’t have to deal with any duplication. You can always work on the same document, and that history is automatically captured behind the scenes,” says Emre.

Paige and Radbee developed detailed specifications for the system upfront to ensure the solution would be a good fit. RadBee delivered a high-quality extension that went above and beyond the original spec.

Results

An almost fully automated system of MDR index tables that aces audits and enables Paige’s team to focus on more high-value work — plus ongoing projects with Radbee to continue iterating and improving the quality system.

With Radbee’s help, the new MDR index has ensured a faster, easier path to iron-clad compliance.

“We’ve been audited by FDA and different certification bodies, with no problems,” notes Emre.

The work completed in April 2021 became a turning point in Paige’s development process. With a system that supports Emre’s topline objectives, the work his team and he does has been significantly up-leveled.

While the original specifications were detailed, Emre found that Radbee went beyond even what he envisioned to solve the problem — and continued to add automated features that make Paige team members’ lives easier.

“RadBee created a few functionalities that we hadn’t thought of or spec’d, such as a “current version” toggle. Rina told us, ‘By the way, we added some additional features for your convenience.’ They are productive in that regard.”

Thanks to the trust built on the MDR index project, Emre has retained Radbee for more extension customizations and other critical quality system upgrades. The time zone difference (Boston/London) has even been a boon, as Radbee can deploy improvements during off-hours. Emre appreciates the flexibility and comfort level, with kind words about the working relationship with Rina and her team.

“Radbee is more than a contractor to us — they are highly embedded in our infrastructure so that they can work with us as fast as possible. This gives us tremendous leverage to work with them on quick decisions and an ad hoc basis without necessarily formalizing every single thing. To me, that’s valuable.”

Agile medical device teams need a QMS that supports a high release cadence. Customizing Confluence to suit your unique needs is an impactful and cost-effective path to greater productivity and better results.

Find out how we can help you automate and otherwise make your QMS system more efficient to keep your medical device releases flowing, so your team can focus on high-value work for happier customers and a healthier bottom line.