Dagmara Kulis, Translation Team Lead, EORTC


This post is the second in our larger series, Industry Insider Interviews. Read our previous post, featuring Ed Potero of Biogen, here.

Tell me about what you do at EORTC, and how long you’ve been in your role?

DK: I am a translator by training with a Master’s degree in English, French and Spanish literature and a Master’s in Translation Studies. I joined EORTC in November 2009. With the quality of life measures created by the EORTC Quality of Life Group, we have a big portfolio of measures that need to be translated. All questionnaire translations, whether academic or commercial, are coordinated by our team. We are involved in the process and review the different steps and the reports to make sure they are completed according to our process, and the quality is up to standards.

We are quite established with our procedure. Thanks to the over 30 years of expertise within the Quality of Life Group, EORTC’s translation procedure was one of the processes that formed the basis for the ISPOR Task Force guidelines. We make sure every translation follows the same steps, sometimes with extra proofreading done by some of our vendors to whom we subcontract our projects.

Last year, you updated your translation process. Was this based on evolving best practices, or in response to the needs of the life sciences industry, or the globalization of clinical trials?

DK: It was mainly to make it more streamlined and describe it in a clear way. We also added the proofreading step that we had been doing anyway. It was a needed update. We have not noticed a lot of changes in the way translations are done. If anything, people do less rather than more because of the timelines. We often receive requests to skip a step but that’s not something we allow.

When people come to you for translation, do people have the expectation that the process can be amended or shortened?

DK: Yes. Timelines are an increasingly important issue. While I thought that understanding of the importance of the process would grow over time, nowadays everyone always wants to make it faster, and people often have problems accepting the minimal 12-14 weeks for new translations to be finished. That’s really the minimum time – and that’s if everything goes well. That includes recruitment for pilot-testing, which can be challenging with all of the data protection issues and the prevalence of rare cancers. It’s a challenge to find enough patients, and that affects the timelines.

You have a process that’s well established – a well oiled machine. Do you see opportunities on the“other side for improvement, from sponsors and providers?

DK: Yes. Communication is often an issue. We’re expected to answer on the same day, but sponsors and other parties may not respond for a week or two when we need something from them in order to finalize the contract and administration. Also, they may send a finalized contract after all this time with an expectation to have a translation done in 2 weeks. This generally does not happen with companies who have dedicated staff for eCOA/ePRO, who handle licensing and translations. We have a good working relationship with them, especially since most of them do a lot of studies and work with us all the time. We see more problems with companies who don’t have dedicated ePRO/eCOA experts, or who have not typically incorporated quality of life measures in their studies and are not aware of how things work.

In general, people in big companies with specialized departments for everything don’t always recognize we are a relatively small non-profit. The commercial world of big corporations is very different from not-for-profit organizations.

Looking at new translations, are there any trends you’ve noticed regarding specific countries or languages?

DK: Yes, we definitely notice trends. For new translations, we always have a batch of translations that are available from the start because, following the EORTC Quality of Life Group’s guidelines, our questionnaires are developed in a number of countries at all stages. We have at least 10 translations once the questionnaire becomes available for external users. Then we see the requests for new translations. There is definitely a rising need for African languages, especially the South African languages, as well as for South East Asia and Asia in general – all of which we did not previously do with the same frequency.

Do you have a position on universal translations for languages that have many different dialects?

DK: Certainly, I even gave a presentation on this topic some time ago. There are definitely conflicting views, depending on who you talk to. Of course, translation vendors will want to have separate versions for every single country because that’s what brings them business. For us, it makes everything more complicated if we have too many local versions to manage. We don’t want 20 different versions of Chinese that don’t differ that much. We want to have a range of languages that are really needed and try to keep the differences to the absolute minimum by harmonizing as much as we can. We do have many different Spanish versions in Latin America because it is hard to harmonize them all together. It can be a political question, where some countries prefer to have their own version, clearly stating it was developed for them locally. It also depends on what the project is. If we have a big project for one sponsor for 3 different Spanish versions, we’ll try to pilot-test and harmonize that version in those 3 countries. However, if we have an existing version for, let’s say, Argentina and then 2 years later have to make a new one for Mexico, we’re not always able to harmonize them because of the time that passed and the fact that they were not done together. Still, within Europe, we try to keep it down to the absolute minimum, so, for example, we have one French version that covers France, Belgium and Switzerland. When we develop questionnaires in French, we always make it a point to have one translator from France and another from Belgium. The translation is also pilot-tested in both countries to make sure it works for both populations. It is the same for Dutch: we have one version of Dutch that works in the Netherlands and Belgium. However, there are some developers who do have 2 different Dutch versions.

You mentioned challenges with patient population availability. What do you do when you can’t find enough patients?

DK: Normally we request 10 patients for the interviews. Our plan B in case of difficulties is to have at least five patients – of course, the more the better. Then, the remaining interviews are done with healthcare providers who work with that population of patients – doctors, nurses, psychologists, etc., as well as laypeople. In this scenario we interview 5 patients, 3 healthcare providers and 2 laypeople who confirm the comprehensibility of the questions. In general, our questionnaires are not very medical, so a lay person should not have a problem formulating an opinion about the translation. Of course, it is better to have patients because they bring firsthand knowledge of their disease, but that’s our plan B.

Are there any other area on this topic that you would like to comment on?

Regarding ePRO related issues, we observe a big increase in licenses for ePROs. More than 50% of our licenses now are for an ePRO version. We are working on having the materials necessary for ePRO migration available in other languages, but that will take time. There are many small things, when you think about standard questionnaires and how, for example, the paper versions say “circle the number”, and the electronic versions should say “select the number”. Also the increasing number of reviews is a challenge we have to address.