Hunnect Outcomes - What are the additional challenges of linguistic validation compared to standard translation projects?
It is well known that an “average” translation project consists of the following steps: translation, editing, proofreading (TEP), and it may also include some additional post-production stages such as reviewing the ready-to-print PDF. These are all well-rehearsed, routine-based tasks that are familiar to any language professional.
When it comes to linguistic validation, although there are some overlaps with translation, the whole process is new and unfamiliar to those that are used to working on traditional translation projects. Linguistic validation projects are longer and have a more complex structure, which means that linguists need to be trained, and the clients needs must be fully understood.
An average translation project involves 4-5 people before it is ready to print or go online. By contrast, linguistic validation can require 12-13 people working on the text to reach its final form.
Step by step
Let’s take a look at what is different about these projects.
When the translatable file arrives, the first step is not to send it to the translator right away, but to have a language expert spend some time on correcting any errors in the source file. We can already ask the client questions at this point if something is not clear. Also, at this stage the files are prepared so that the linguists can work on them in a CAT tool.
The source is then sent to two separate translators, unlike in usual projects with a single translator.
This is followed by the reconciliation step. At this point, another linguist works with both translations and chooses the better of the two translations for each segment, or may occasionally formulate a third version.
The reconciled file is sent to the editor, and then to the linguist who carries out the monolingual proofreading.
Then comes the back translation, after which, in theory, the file should be perfectly equivalent to the source, at least in terms of meaning. The back translation is compared with the source file by another translator, and if any discrepancies are found, they should be reported to the lead linguist so that the latter can consider whether to change the reconciled translation.
If all goes well, there should be no translation errors left at this point.
Then the monolingual translation file is reviewed by a specialist, preferably a physician specializing in the field of the given clinical trial, who may make suggestions for changes, which will also be considered by a linguist. In this clinician review step, the project manager should always take into account that the physician is not a linguist. For a successful collaboration, this must be remembered when setting the deadline and compiling the instructions, because this specialist primarily works with patients.
One last but important task
Interviewing patients (also called cognitive debriefing) is a challenging task in many ways, as it is quite different from standard linguistic tasks. However, it is at this stage that it becomes clear whether the text would be appropriate for the target audience.
Here, the interviewer – who may be the project manager or an external subcontractor – contacts the patient via the communication channel specified by the client, and reviews the translated text together with the patient. The main goal of this step is to eliminate any structures or terms that the end user of the text may not understand easily.
The challenges of this step have been described in a previous blog post.
Writing the report and quality assurance
A very interesting stage of linguistic validation is writing the final report.
The report form to be filled can be very simple or extremely complex. The simplest method is to essentially keep a record of who did what and when during the project, with names and dates. More complex reports can even include interlingual statistical analyses of multilingual projects. Of course, this is less common. In any case, the process of this step should be tailored to the client’s expectations.
Quality assurance is not just verification using software: reconciliation and back translation are also very powerful ways to detect possible mistranslations.
Pricing issues
Pricing for the customer is not easy either, as there are processes that are charged based on the actual time spent. These include the reconciliation, comparison after back translation, harmonization, the cognitive debriefing and writing the report. Furthermore, the exact word count of the back translation becomes known only after the translation has been completed.
The question may arise: why is this long and complex process necessary?
A language service provider performing such projects takes full responsibility for the quality. Not even the slightest mistake is permissible, because what is created will end up in the hands of the study subjects. Incorrect or ambiguous translations can, in the worst cases, reduce the effectiveness of the entire clinical trial or distort the data. The target audience is very special, as the readers and users of the target text can be people from a wide range of educational and socio-cultural backgrounds, so the text must be perfectly comprehensible, simple and unambiguous.