The number of global trials grows every year, with study teams reaching out to increasingly diverse patient populations. These broader populations bring a growing variety of languages, cultures, and experiences – all of which need to be considered when recruiting or interacting with clinical trial participants. While regulations vary from country to country, translation of clinical trial materials into the local language allows participants to understand what the trial involves and follow directions.
Patient and clinical trial materials are continuously adapting to better reach patients and populations with increasingly imaginative and interactive recruitment strategies. For example, activities like crossword puzzles and interactive games have now become common tools for better engaging patients.
With the volume of translations growing and the demands on the translators getting more complex, what role should machine translation play in the translation of clinical trial materials?
What is machine translation?
The American Translators Association defines machine translation as the use of automated software that translates text without human involvement. This technology is not to be confused with computer-assisted translation tools.
Machine translation came to fruition in the early 1950s. Today, we see the rise of well-known apps and free web machine translation tools such as Google Translate and Microsoft Translator. These apps provide generic machine translations and have become a household tool. However, these tools differ greatly from customizable machine translation, which is taking hold within the translation industry. Customizable machine translations are believed to be more accurate. They focus on field-specific terminology, for example, within the law or medical terminology. Machine translations can even be adapted to focus on client-preferred language or terms.
Over the years, machine translation has advanced, thanks to technological developments. The latest phase of neural machine translation is proving to be the most promising, providing the most fluid and accurate translation results to date. But the technology is still an emerging technology. Like most – and possibly all technology — machine translation has its benefits and pitfalls.
- Increased productivity. Machine translation capacity far outstrips any professional human translator. This can result in vastly reduced timelines, especially when handling large translation texts.
- Cost saving. With the increase in productivity comes a decrease in costs, as less time and fewer vendors are required to fulfill a translation request.
- Enhanced consistency in translation across large amounts of text. This can also enable consistent use of client-preferred language in translation that may not be possible with a multi-human translation team.
- Machine translation stumbles when it comes to cultural nuances. Machine translation still does not have the capacity to handle cultural nuances or slang in translation as a human native translator would.
- Lack of standardization. There is a wide range of software available. Developers on the frontier of machine learning are trying new approaches with varying and somewhat unpredictable results.
- Less reliability with languages spoken by small numbers of people. Machine translation relies on large amounts of bilingual text or data. With languages of limited diffusion or rare languages, machine translation struggles to output as sufficiently as with common languages.
Machine translation of clinical study materials plus the human touch
Best practices indicate machine translation requires post-editing, and depending on the text, pre-editing by a qualified human translator to deliver an optimal translation product. Pre-editing by a human translator ensures nuances, such as changing Fahrenheit temperature to Celsius, are tackled ahead of machine translation. Post-editing enables the translator to review the translation output. This ensures translation accuracy and enhances the fluidity of the final product.
It is undeniable that machine translation has vastly progressed since its conception. It is a great aid if utilized appropriately. However one would argue that for complex translation, including needs applicable to the translation of clinical trial materials, it is not yet capable of being used as a single source.
Post-editing and potentially pre-editing are vital to ensure the delivery of the optimum product. As for translators, with the rapidly increased use of machine translation and other technology we have seen and foresee in the future, their skillsets are ever expanding and evolving. This will forge a stronger collaboration between humans and technology in the field of translation. It will be fascinating to see these developments.