Life Sciences
Forward Translation, Back Translation and Cognitive Debriefing for Clinical Outcome Assessments
Linguistic validation of a clinical outcome assessment (COA) usually follows the 10-step ISPOR process (Wild et al., 2005). The steps are preparation, two forward translations, reconciliation, back translation, back translation review, harmonization, cognitive debriefing with 5 to 8 respondents, finalization, proofreading and a final report. FDA reviews this process when COA data support labeling claims.
- By Umar Khan
- October 5, 2026
- 6 min read
Key takeaways
- FDA expects sponsors to document how a COA was translated and culturally adapted.
- Most validated COAs are copyrighted, so licensor approval usually comes first.
- The ISPOR process uses two forward translations, reconciliation, back translation and cognitive debriefing.
- Use back translation as one check, alongside expert reconciliation and cognitive debriefing.
10 steps
In the ISPOR linguistic validation process
5 to 8
Debriefing respondents per language
2009
FDA PRO guidance on translated measures
Why Does FDA Care How a COA Was Translated?
Many trials pool patient-reported outcome (PRO) data across countries and languages. Pooled data are only valid if every language version measures the same concept in the same way.
FDA’s 2009 guidance, “Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims,” addresses this directly. It recommends that sponsors show that content validity and other measurement properties are “adequately similar between all versions used in the clinical trial.” FDA states, “We will review the process used to translate and culturally adapt the instrument.”
The guidance appendix lists what to submit:
- The translation and cultural adaptation process.
- A description of patient testing and language-specific decisions.
- Copies of translated versions.
- Evidence that measurement properties are comparable.
FDA’s Patient-Focused Drug Development (PFDD) guidance series builds on this. Guidance 3, “Selecting, Developing, or Modifying Fit-for-Purpose Clinical Outcome Assessments,” was issued in October 2025. It lists translation from one language to another as a type of COA modification. It recommends “using best practices for adaptation/translation” and cites Wild et al. (2005), the PRO Consortium process (Eremenco et al., 2018) and McKown et al. (2020).
Who Must Approve a COA Translation?
Most validated COAs are copyrighted. Mapi Research Trust explains that to use a COA in a clinical trial, “whether it is the original version, a translation and/or an electronic version (eCOA), sponsors must secure permission to use it with a study-specific license.”
New translations usually must be approved by the licensor on behalf of the developer. The licensor may also supply the developer’s translation guidelines. Check for an existing validated translation before commissioning a new one.
What Are the 10 Steps of Linguistic Validation?
| Step | What Happens | Output |
|---|---|---|
| 1. Preparation | Obtain permission, define concepts for each item and recruit a key in-country person. | License, concept definitions |
| 2. Forward translation | At least two independent translators, native in the target language, translate the instrument. | Two forward versions |
| 3. Reconciliation | The forward versions are merged into one. The project manager reviews decisions. | Reconciled version |
| 4. Back translation | At least one translator, native in the source language, translates the reconciled version back. | Back translation |
| 5. Back translation review | The back translation is compared with the original. Discrepancies are investigated and the translation is revised. | Revised translation |
| 6. Harmonization | Language versions are compared with each other and the original for consistent concepts. | Harmonization notes |
| 7. Cognitive debriefing | The translation is tested with 5 to 8 native-speaking respondents in the target country. | Interview findings |
| 8. Review and finalization | Findings are compared with the original, and agreed changes are made. | Final translation |
| 9. Proofreading | Spelling, diacritics, grammar and formatting are checked. | Proofread version |
| 10. Final report | Methods and every item-level decision are documented. | Final report |
ISPOR presents these steps as “practical guidance rather than prescriptive rules.” The PRO Consortium’s 2018 process builds on ISPOR with 12 steps. It keeps back translation as a required step and calls for a back translator who has not seen the source questionnaire.
How Does Cognitive Debriefing Work?
Cognitive debriefing tests whether real respondents understand the translation as intended. ISPOR lists four purposes:
Assess comprehensibility and cognitive equivalence.
Test wording alternatives the translators could not resolve.
Flag items that may be inappropriate at a conceptual level.
Identify other sources of confusion.
ISPOR recommends 5 to 8 respondents per language in the target country. Respondents should be native speakers who represent the target population by sex, age, education and diagnosis. The PRO Consortium sets a minimum of 5 participants per language and country.
In a typical interview, the respondent completes the questionnaire. The interviewer then asks the respondent to explain items in their own words and to describe any difficulty. The project manager reviews the findings with the key in-country person. Changes are made only where respondent comments justify them.
Planning linguistic validation for a global trial?
We manage forward and back translation, reconciliation and cognitive debriefing with full documentation.
When Is Back Translation Useful, and When Is It Not?
ISPOR describes back translation as a quality control step. It shows whether “the same meaning” comes through when the text returns to the source language. ISPOR suggests a more literal back translation for medical symptoms and a more conceptual one for subjective items, such as quality of life.
Back translation is useful when:
- The developer, licensor or sponsor reviewer does not read the target language.
- You need a documented check for omissions or changed meaning.
- The licensor or sponsor procedure requires it.
Back translation has limits. It does not show whether the target text reads naturally. It does not test patient understanding. A smooth back translation can hide awkward wording in the target language.
The evidence is limited. Epstein et al. (2015) created four English to French translations of the Health Education Impact Questionnaire, with and without back translation or an expert committee. Bilingual assessors ranked the translations reviewed by a committee highest. The authors concluded that back translation had “moderate impact,” while the expert committee helped ensure accurate content.
Use back translation as one check, alongside expert reconciliation and cognitive debriefing.
What Are the Most Common Pitfalls?
Starting translation before the license and developer guidelines are in place.
Skipping concept definitions, which leads to literal translations.
Using a single forward translation.
Shifting recall periods, frequency terms or intensity words.
Recruiting debriefing respondents who do not match the trial population.
Skipping harmonization. ISPOR warns that inconsistent language versions may make it difficult to aggregate data.
Writing a thin final report. ISPOR warns that poorly documented translations may not be usable.
Not checking the final text again after it is placed into eCOA screens or paper layouts.
Frequently Asked Questions
How many participants are needed for cognitive debriefing?
The ISPOR task force recommends 5 to 8 respondents per language, tested in the target country. The PRO Consortium’s 2018 process sets a minimum of 5 participants per language and country. Respondents should match the target population for as many criteria as practical, including sex, age, education and diagnosis. FDA’s PFDD Guidance 3 adds that sponsors should include respondents who reflect the range of literacy and numeracy in the target population.
Does FDA require back translation for COAs?
FDA’s 2009 PRO guidance does not prescribe specific translation steps. It states that FDA will review the translation process and expects evidence that language versions are comparable. FDA’s PFDD Guidance 3 points to published best practices, which include back translation. Licensors may also set their own translation requirements.
Can we use an existing translation of a COA?
Often, yes. Many copyright holders and licensors distribute validated translations under a study-specific license. Using an existing translation can save time and supports comparability with earlier studies. You still need permission from the licensor. Confirm that the translation fits your target country and population before you use it.
Does linguistic validation apply to ClinROs, ObsROs and PerfOs?
Yes. McKown et al. (2020) published good practices for translating clinician-reported, observer-reported and performance outcome measures. FDA’s PFDD Guidance 3 cites this paper alongside the ISPOR and PRO Consortium methods. For PerfOs, FDA notes that cognitive interviews about task instructions, combined with pilot testing, can confirm that patients understand the task.
How AsianText Helps
Documented Linguistic Validation for Global Trials
AsianText supports sponsors, CROs and eCOA providers with COA and PRO translation. Our ISO 17100 certified workflow covers independent forward translations, reconciliation, back translation and proofreading by native medical linguists in 200+ language combinations. We follow your licensor’s requirements and document each decision for your final report. Our localization and multilingual desktop publishing teams check eCOA screens and paper layouts. See our life sciences expertise. Contact us to scope your COA translation project.
This article provides general information, not legal advice. Last updated October 5, 2026.
Written by Umar Khan
Part of the AsianText team, writing about language access, translation and interpreting for government, healthcare and life sciences.
Sources
- Wild D, et al. Principles of Good Practice for the Translation and Cultural Adaptation Process for PRO Measures. Value in Health, 2005 (ISPOR)
- FDA: Patient-Reported Outcome Measures guidance (December 2009)
- FDA: PFDD Guidance 3, Selecting, Developing, or Modifying Fit-for-Purpose COAs
- FDA: Patient-Focused Drug Development Guidance Series
- Eremenco S, et al. PRO Consortium Translation Process. J Patient Rep Outcomes, 2018
- Epstein J, et al. Cross-Cultural Adaptation of the Health Education Impact Questionnaire. J Clin Epidemiol, 2015
- Mapi Research Trust: COA Copyright (July 2023)
Related Services
Put This Guide Into Practice
Translation
Vital documents, notices and technical content, with certified translation and three quality checks.
Life Sciences
Translation and linguistic validation for clinical research and study documents.
Plain Language & Community Review
Clear English first, then paid community reviewers who confirm the message lands in every language.
Translating for a Clinical Study?
Tell us the languages, the content and the deadline. We reply with a fixed quote within one business day.