Santium is a specialized translation and data collection company that focuses on validating clinical outcomes assessments for clinical research and practice and research.

Our services include linguistic validation, and are optimized for all deliverables to meet regulatory standards and produce reliable and culturally-sensitive data.

Our work, entrenched in psychometrics, involves repurposing existing patient-reported outcomes (PROs) for new respondent groups and translating, culturally adapting and validating clinical research materials.

We produce translations, administered to adults and children from diverse populations, that are linguistically accurate and culturally appropriate. This is particularly crucial in CNS trials, where understanding the nuances of language, culture and local education guidelines can significantly impact the reliability and validity of clinical data.

To support these core services, we also offer a range of complementary services such as graphic design, desktop publishing, medical editing, scientific writing, and we translate all types of other content.

For example, manuals, protocols, informed consent, textbooks, videos, course content, instructions, presentations, software applications, website content, etc.

We invite you to contact us at info@santium.com to learn more about how we can support your translation project.

Santium packs deep expertise in validating psychometric questionnaires, hence understanding the process of (neuro)psychological test development, and rules for designing and measuring specific states, traits, functions and behaviours.  This becomes critical during translation intelligence tests and other neuropsychological batteries.

Through our specialized focus, multi-disciplinary blend of translation, clinical and editorial specialists, we easily customize our services to conform to any study design, regardless of the study size and budget.

Key features that set us apart are:

• Expertise in psychometric measurement
• 15+ years’ experience translating and validating CNS assessments and batteries
• Specialized hybrid cognitive debriefing process for children and adults
• Focus on linguistic and cultural sensitivity
• Experienced team with relevant expertise to study design and choice is assessments

Translation and linguistic validation are two related methodologies but have distinctly different scopes, processes, outcomes and uses.

Linguistic Validation consists of Translation and Cognitive Debriefing.  Cognitive Debriefing is the “validation” part of the Linguistic Validation process.  If Cognitive Debriefing doesn’t take place, the translation is not linguistically validated; it’s just a translation.

Scope:

Translation focuses on accurately conveying content from one language to another, while linguistic validation includes translation but also addresses cultural relevance, understanding, and equivalence across languages.

Process:

Translation typically involves one or two translators and possibly a reviewer. Linguistic validation involves multiple steps, including translation, back-translation, linguistic and expert review, and cognitive debriefing with target responders.

Outcome:

The outcome of translation is a text that accurately reflects the source material in another language. The outcome of linguistic validation is a culturally and linguistically adapted, comparable/equivalent text that is understood in the same way by its target language groups.

Use:

Translation is used for a wide range of documents, including manuals, training content, information booklets, and textbooks. Linguistic validation is specifically used in Life Sciences fields, where it’s crucial that instruments like clinical outcomes measures work consistently the same way across different languages and cultures, despite their unique syntax and grammatical structures, and behavioural norms.

Artificial Intelligence (AI) has been gradually modifying the linguistic validation workflow.

While AI is improving, it still lacks the ability to fully grasp context, cultural nuances, and the subtleties of language that are critical in linguistic validation. Hence, human oversight is still essential to ensure accuracy and appropriateness.

As a result, we do not allow the use of AI technologies in forward translation. This also applies to MT translations with post-editing.

We will periodically revisit the current process as AI technology evolves.

Use Case Example

Though AI does not replace human expertise in the forward translation processes, it does offer technological benefits that enhance the efficiency, accuracy, and consistency of the validation process, and nearly eliminating the need for a human backward translation.

Scope:

AI-powered technologies are used to generate backward translations during the linguistic validation process.

The key purpose of a backward translation is to help identify errors or inaccuracies by checking for semantic equivalence between the original and translated content, and to ensure that the translation conveys the same meaning as the original, without any distortion, loss, mistranslation, localizations that alter the intended meaning, or addition of meaning.

This provides a speaker of the source language with an opportunity to conduct a translation quality pre-check and return the translation to the translator(s) for corrective action, before giving an expert reviewer access to a translation with known errors, and unnecessarily complicating and extending review time.

Risk:

It is critically important that the backward translation is never aesthetically edited, because doing so could unintentionally disguise even the slightest distortion in meaning and inaccuracies. 

A backward translation reveals to a source language speaker whether the translation is not only linguistically correct but also culturally and contextually appropriate. It helps ensure that the translation makes sense in the cultural context of the target audience while still retaining the meaning of the original text.

For less complex clinical outcomes measures and other documents, a MT-generated backward translation has also proven to be a reliable quality pre-check tool during the translation process.

For more complex content, it is still better for human translators to produce backward translations.

None.

The 2009 Patient-Reported Outcome (PRO) Guidance for Industry, issued by the U.S. Food and Drug Administration (FDA) only briefly touches on the importance of translation and cultural adaptation when patient-reported outcomes measures are used in international clinical trials.

The guidance is primarily focused on ensuring the reliability and validity of PRO measures.  Although the guidance is specific to PROs, its content can be readily applied to any type of a clinical outcomes measure (e.g., ClinROs).

Not including a translation methodology in the PRO guidance was intentional, for two key reasons:

1. FDA chose not to limit researchers to a single prescribed method of translation

2. The industry was in the process of testing a linguistic validation method developed by a working group representing the International Society for Pharmacoeconomic and Research Outcomes (ISPOR), published it in 2005 (Wild et al., 2005), and at the time still in its infancy stage of adoption.  This ISPOR-developed method eventually became the gold standard that is now known and accepted throughout the industry.

FDA’s decision not to prescribe a translation method has had a dual effect across the Life Sciences industry: 

1. Cognitive debriefing is often omitted from “linguistic validation” projects, mainly due to longer project timelines that are necessary to complete the pilot test. Therefore, a very high percentage of translations used in clinical trials are not linguistically validated.

2. Technological advances in the translation industry made certain steps in the methodology, including its variations, redundant and unnecessary.

After 15 years of selective application, the linguistic validation method is due for a practical revision that must include processes for adapting untranslatable content.

Not true.  

Let’s review what “translation equivalency” actually means:

Translation equivalency refers to the degree to which a translated content accurately reflects the meaning, intent, and impact of the original source text while being culturally appropriate and understandable in the target language.

Given the prescribed meaning ans purpose of “translation equivalency”, the wording of items in an outcomes measure, or a survey, may have to change to accurately convey a particular concept to the target responder, especially when the original concept is no longer relevant to the target responder, or to the context of the responder’s culture or environment.

In some cases, the concept may not exist in the target country altogether, or may only be applicable to other types of responders.

Therefore, when translating clinical outcomes measures, translators and study project managers must unlearn this misconstrued resistance to adapting psychometric instruments.

It will be crucial to validate the translation, at least via thorough cognitive debriefing, before using it in its intended context.

If a user’s manual or any form of instruction accompanies the instrument, or survey, ensure that changes made in the questionnaire(s) are reflected in the administration instructions, including screenshots of sample scoring forms, demo items, etc..

Likely true, if done by someone unqualified to make that kind of a change.

This unfortunately includes almost all translators, hence the responsibility to provide clear and accurate instructions on how to do it, including any necessary training, lies with the party assigning the job to the translator.

The validity of a rating scale refers to how well the scale measures what it is intended to measure. Therefore, from a psychometric standpoint and working with translations, replacing and rewording items is an essential part of preserving the translated instrument’s validity.

A translated clinical outcomes measure, even when flawlessly translated to mirror the original source, can instantly lose its validity when some or more of its items are not relevant to the target language speakers and their culture.

For example, items that measure stigmatized concepts, verbal fluency, adaptive behaviours, language development milestones, house chores, or country-specific knowledge across tiered age groups, just to name a few.

Whenever changing source text becomes necessary, it’s crucial to revalidate the scale to ensure that it still measures the intended construct validly and reliably.

More often than not, it’s the author’s representative or agent, or internal corporate policies governing the representative or agent, that don’t allow for the changes to be made; it’s usually not the author.

If the author psychometrically validated the instrument, the author will understand the need for the change, assuming that you have a valid argument to support the change. Also, the author must know that only a qualified individual, or a group, will be making the change. 

If you have a strong case, the author will either make the changes for you, or will allow you to make changes to source text within specific parameters and/or with the author’s guidance. 

Depending on the nature of the change, you may also need to adjust the scoring algorithm.

If the author, or the author’s representative or agent, refuse(s) to allow you to change source text, especially when you provide a solid psychometrically-driven argument for the change, your project or study may be better off with another instrument in its assessment schedule lineup.

Whether cognitive debriefing is necessary depends on several factors, including the purpose of the translation, the complexity of the content, and the importance of ensuring that the translated content is understood as intended. A thorough translatability assessment will assist in making this decision.

Cognitive debriefing is highly recommended when translating or adapting clinical outcomes measures for use in different languages or cultural contexts.

When Cognitive Debriefing is Necessary

Importance:

1. In clinical trials that involve patients from different countries and/or linguistic backgrounds.

2. To ensure that patients understand the questions about their health, symptoms, or quality of life in the way the author(s)/developer intended.

3. In surveys or questionnaires where the results will inform critical decisions (e.g., public health surveys, expert consensus guideline surveys, large-scale research studies).

4. When adapting content to fit a new cultural context or a different patient group.

Goals:

1. To identify and resolve any issues with comprehension, ensuring that patients interpret the questions in the same way, regardless of their language or cultural background, which is essential for the validity and reliability of the trial’s data.

2. To maintain the conceptual equivalence of the measure and ensure that it accurately captures the patient’s experience in the target language.

3. To ensure cultural relevance and appropriateness while maintaining the original intent of the instrument, and that culturally specific terms, idioms, and concepts are appropriately translated and understood by the target population.

When Cognitive Debriefing May Not Be Necessary

1. Simple, Direct Translations:

Importance: Straightforward texts that do not involve complex concepts or culturally specific content.

Example: Translating a user’s manual for outcomes measure where the instructions are clear and universally applicable.

Risk: The risk of misinterpretation is low, and the translation can be verified through standard proofreading and review processes.

2. Internal and Regulatory Documents:

Importance: For internal and regulatory documents where the audiences are familiar with the content and context.

Example: Internal training materials for a multilingual team that already understands the context and terminology.

Risk: The potential impact of minor misunderstandings is limited.

3. Non-Critical Communication:

Importance: For communications where slight variations in understanding are acceptable.

Example: Casual marketing materials or social media posts where the exact wording is less critical.

Risk: The stakes are lower, so precise alignment between the source and target texts is less critical.