Consider these:

Trying to read each of these is very irritating, isn’t it? In some cases, you cannot perceive where one letter ends and the other starts.

Yet this is the same kind of difficulty that patients around the world encounter when faced with fonts and layouts more suited to Latin scripts like English or French. This is something that is frequently overlooked: the assumption that fonts that work well in the source language work just as well for non-Latin scripts, like Georgian, Arabic, Tamil or Chinese. However, this assumption can lead to serious legibility and usability issues. This is especially pertinent for Clinical Outcome Assessments (COA), where precision, clarity and patient comprehension are critical.

Eye-tracking studies and user experience research repeatedly show that script density, character structure and line spacing play a vital role in how easily text can be read. When this reality is ignored, it has repercussions: it does not only slow down reading, it can also compromise the validity of patient-reported data.

One-Size-Fits-All Fonts?

The majority of COAs originate in a variety of English and use very familiar Latin fonts, such as Arial, Calibri, Helvetica or Times New Roman. These are specifically designed for optimising the visual processing of light, unconnected characters and a left-to-right reading direction. When COAs are translated into languages with more complex scripts but retain the original font and layout, then problems arise: they can have a similar visual effect as with the English examples I provided above.

This article explores the challenges presented by typographic mismatches across 10 global scripts and how we can fix the issue.

1.       Arabic

Challenge: Arabic is a cursive script, where letters connect to form a flow of words. Letter shapes change depending on their position: beginning, medial and end.

Problem: Latin fonts such as Arial and Calibri are poorly read in Arabic. They are too geometric and disconnected for the language, making words seem fragmented and unnatural.

Impact in COAs: Patients might struggle to track connected words in a form where the font breaks the flow, especially in small print.

Better fonts to use: Noto Naskh Arabic, Droid Arabic Naskh.

2.       Hindi (Devanagari script)

Challenge: The Devanagari script features horizontal headstrokes and stacked consonants.

Problem: Fonts with wide spacing like Verdana can cause the script to seem broken or overly spaced, disrupting the reading rhythm.

Impact in COAs: Headings or questions may seem disjointed and this increases reading time and cognitive load.

Better fonts to use:  Noto Sans Devanagari, Lohit Devanagari

3.       Tamil (Agglutinative Script)

Challenge: Tamil has looping, rounded letters and long compound words.

Problem: Sans Serif fonts can create excess white space due to loose line spacing, breaking up the Tamil text, which needs to be compact

Impact in COAs: Compound items or shared stems in questionnaires become visually fragmented, risking misinterpretation.

Better fonts: Latha, Noto Sans Tamil, Bamini.

4.       Thai (Tonal Script)

Challenge: Thai has stacked vowel marks and tone markers above and below base letters.

Problem: Font with tight white line heights, such as Arial, can cause diacritic collision, making the text hard to read.

Impact in COAs: Critical instructions or response choices might be visually cramped or illegible.

Better fonts: Noto Sans Thai, Sarabun, TH Sarabun New

5.       Georgian

Challenge: Georgian uses dense, rounded letterforms with no upper or lowercase distinction.

Problem: Fonts like Futura flatten Georgian’s natural rhythm, making text harder to follow.

Impact in COAs: List items or response scales lose clarity, especially when laid out like an English form.

Better fonts: Sylfaen, Noto Sans Georgian, BPG Nino Mtavruli

6.       Chinese (Logographic)

Challenge: Chinese characters are dense, square units with multiple strokes

Problem: Thin fonts like Arial make detailed characters look faint and hard to distinguish, especially at small sizes.

Impact in COAs: Instructions and answer choices may lose contrast and clarity, increasing reading time and visual fatigue.

Better fonts: Source Han Sans, SimSun.

7.       Japanese (Mixed Script)

Challenge: Japanese readers switch between kanji (logographic) and kana (syllabic) scripts within the same text.

Problem: Fonts like Times New Roman break the natural flow between these scripts, disrupting the reading rhythm.

Impact on COAs: Key questions and responses may feel visually fragmented. This reduces fluency and comprehension.

Better fonts: Meiryo, Noto Sans JP

8.       Korean (Hangul)

Challenge: Hangul syllable blocks combine consonants and vowels into compact square shapes

Problem: Fonts with wide spacing and Sans fonts, distort Hangul’s balance and visual rhythm

Impact on COAs: The distorted alignment of syllable blocks may lead to reading strain and slower form completion.

Better fonts: Nanum Gothic, Noto Sans KR.

9.       Urdu

Challenge: Urdu is written in a diagonal and calligraphic script.

Problem: Latin fonts disrupt the flowing, slanted structure of the script, making text feel unnatural.

Impact on COAs: Patients may find the translated text harder to engage with, increasing the cognitive burden.

Better fonts: Noto Nastaliq Urdu, Jameel Noori Nastaleeq.

10.  Khmer

Challenge: Khmer has tall, looping characters with stacked diacritics.

Problem: Fonts with tight line heights like Arial and Roboto cause diacritics to collide or overlap.

Impact in COAs: Dense text blocks may become visually cluttered, leading to misinterpretation of items.

Better fonts: Battambang, Noto Sans Khmer.

What Does This Mean?

Latin fonts, optimised for lightweight, spaced-out letters, do not work for dense or connected scripts. Applying them without adaptation creates:

• Reduced legibility

• Visual clutter

• Patient disengagement.

In clinical settings, this compromises data quality, especially when patients are under time pressure, fatigue or stress.

Typography needs to be functional and legible, not cosmetic.

In COA development, typography requires early-stage planning:

• Choose fonts suited to the density of the script and reading flow.

• Allow flexibility of layout post-translation

• Consider eye-tracking data to guide design decisions during development.

Without this, translated COAs may be technically accurate in translation, but visually unreadable in practice. It also highlights that this really does need to be addressed during development. Once validated, layout change is not considered. It is a shame that regulations around COA are geared towards one-size-fits-all.

Thank you for reading,

Mark Gibson

Leeds, United Kingdom, March 2025