01
Executive thesis
Translation asks whether the Arabic says the same thing as the English. Parity asks whether the Arabic-speaking patient can complete the same decision with the same confidence. These are different questions and they produce different work.
A site can be translated with complete accuracy and still fail. The Arabic page says exactly what the English page says, yet it targets terms nobody searches, omits the practitioner detail that was added to the English profile last quarter, and ends at a booking flow that reverts to English. Every sentence is correct. The patient still cannot finish.
In Saudi Arabia this matters more than in most bilingual markets, because the Arabic-speaking audience is not a minority segment being accommodated. For most providers it is the majority of the addressable patient population, being served by the version of the site that receives less investment, less governance and less review.
The decision this article supports is how to budget and govern bilingual publishing. Arabic needs to be resourced as a parallel product with its own research, its own review cycle and its own measurement, not as a translation task appended to the end of the English workflow.
02
Numbers that frame the issue
| Figure | What it tells us | Type | Source, date and geography |
|---|---|---|---|
| 99% of individuals aged 15 to 74 used the internet | The Arabic-speaking patient population is fully digital, so language gaps are not offset by offline access | External evidence | General Authority for Statistics, ICT Access and Usage, 2025, Saudi Arabia |
| Google recommends separate URLs for each language version, with hreflang annotations | Language handling is an architectural decision with discoverability consequences | Official platform guidance | Google Search Central, multi-regional and multilingual sites, current guidance, global |
The GASTAT figure establishes connectivity, not healthcare search behaviour. Google’s technical guidance supports correct language matching and does not guarantee ranking outcomes.
03
Why translation does not create parity
Four failure modes recur, and they are independent of translation quality.
Search intent diverges. Arabic-speaking patients do not search the Arabic rendering of English keywords. They search the words they use when describing a problem to a family member, which are frequently colloquial, sometimes regional and often mixed with English clinical terms. A page targeting the formal Arabic name for a condition can miss the query volume entirely, while a page built from actual Arabic query research finds it. Translating a keyword list translates the wrong thing.
Depth erodes over time. The English page is the source of truth. It gets updated when a new consultant joins, when a service moves site, when insurance arrangements change. The Arabic page was accurate on the day it was translated. Eighteen months later it describes an older organization, and nobody has noticed because nobody reads it internally.
Trust evidence is thinner. Practitioner profiles are the clearest example. The English profile has scope, sub-specialty, languages and publications. The Arabic profile has a name, a title and a photograph. The Arabic-speaking patient is asked to make the same decision with less evidence.
The journey breaks at the transaction. The content translates. The booking system does not. The patient reads in Arabic, decides in Arabic and then meets an English form, an English error message or an English confirmation SMS. The break lands precisely where confidence matters most.
None of these is fixed by a better translator. They are fixed by treating Arabic as a product.
04
The eight dimensions of bilingual trust
The DEMA Bilingual Trust Parity Framework measures parity across eight dimensions. Score each 0 to 2, where 0 means absent or substantially weaker, 1 means present with gaps, 2 means equivalent in decision value.
1. Discovery parity. Does the Arabic site rank for the terms Arabic-speaking patients actually use, researched natively rather than translated?
2. Coverage parity. Does every service, condition and location page that exists in English exist in Arabic?
3. Depth parity. Do the Arabic pages carry equivalent detail, or a shortened summary?
4. Practitioner parity. Do Arabic practitioner profiles carry the same scope, sub-specialty, languages and credentials as the English ones?
5. Access parity. Are insurance acceptance, hours, referral requirements and site availability stated equally in both languages?
6. Transaction parity. Can a patient complete booking, from entry to confirmation message, entirely in Arabic?
7. Freshness parity. When English is updated, is Arabic updated in the same cycle, or does it drift?
8. Review parity. Does Arabic content receive the same clinical and compliance review as English, by reviewers working in Arabic?
The dimensions are not equally weighted in practice. Coverage and depth parity affect confidence. Transaction and access parity affect whether the patient can complete anything at all, which makes them the two to fix first when resources are limited. Discovery parity determines how many Arabic-speaking patients arrive to experience either.
Sixteen points available. The two dimensions most often scoring zero are transaction parity and freshness parity, and they are the two that most directly cost patients.
05
How Arabic search intent and terminology diverge
Three specific divergences are worth building into research practice.
Register. Patients describe symptoms in everyday language, not in the formal Arabic used in medical writing. The formal term for a condition and the phrase a patient types can be entirely different words. Content written by clinicians defaults to the formal register, which produces pages that are medically excellent and undiscoverable.
Code switching. Saudi patients routinely search using English clinical terms inside Arabic sentences, particularly for procedures, drug names and specialty names. A rigid rule that Arabic pages must contain only Arabic terminology removes the exact strings people search for. The Arabic term should lead, with the English term retained where patients genuinely use it.
Transliteration variance. Names, drugs and procedures transliterate multiple defensible ways. This affects practitioner discovery in particular, and is the same problem described in [The Invisible Specialist Problem: Why Clinical Reputation Does Not Create Digital Visibility]. A written standard, applied consistently, resolves most of it.
| Divergence | What the content team writes | What the patient searches | Consequence |
|---|---|---|---|
| Register | Formal clinical term for the condition | Everyday description of the symptom | Page ranks for a term with little patient volume |
| Code switching | Arabic-only terminology, English removed on principle | Arabic sentence containing the English procedure name | The searched string does not appear on the page |
| Transliteration | One spelling of a practitioner or drug name | Several defensible spellings | Signals split across variants, none of them strong |
| Question form | Statement headings translated from English | Direct questions, often conversational | Content does not match how the query is phrased |
The practical implication is that Arabic keyword research must be conducted in Arabic, by people who speak it as patients do, using Arabic query data. A translated keyword list is not research.
There is a second-order effect worth planning for. Arabic content built this way tends to be more distinctive than its English counterpart, because far fewer competitors have done the work. In service lines where English search is saturated, the Arabic side is often the cheaper route to visibility, and it addresses the larger share of the patient population. That is an argument for sequencing Arabic first in some service lines rather than treating it as the follow-on task it usually becomes.
06
Technical architecture: URLs, hreflang and crawlability
Google recommends keeping each language version on a separate URL and using hreflang annotations to indicate the relationships. For healthcare sites this has consequences beyond the technical.
Separate URLs allow the Arabic version to be updated, measured and governed as an independent asset. They also make drift visible: if an Arabic page has not been updated since 2024 while its English counterpart was revised twice, that is a reportable fact rather than an invisible one.
Four requirements follow:
- Self-referencing canonical tags on each language version, so neither is treated as a duplicate of the other
- Reciprocal hreflang annotations between the pair
- Language-specific metadata written natively, not translated, including titles and descriptions
- Right-to-left handling that mirrors navigation, breadcrumbs and directional elements while keeping numbers, formulas and Latin terms in left-to-right order
The last point is where most implementations produce visible errors. A phone number, a dosage or a URL rendered in reversed order inside an Arabic layout signals carelessness in exactly the context where a patient is assessing whether the organization is careful.
07
A page-pair comparison
The following comparison is illustrative. It describes a composite page pair constructed to demonstrate the scoring. The scores are not measured and are not a DEMA benchmark.
Composite pair. A cardiology service page, English and Arabic.
| Dimension | English | Arabic | Score |
|---|---|---|---|
| Discovery | Targets researched English terms | Targets translated equivalents of the English terms | 0 |
| Coverage | Page exists | Page exists | 2 |
| Depth | 900 words, six conditions listed, three procedures explained | 400 words, three conditions listed, procedures named only | 1 |
| Practitioner | Four cardiologists named and linked, scope stated | Four names listed, no scope, no links | 0 |
| Access | Insurance, hours, two sites stated | Hours only | 1 |
| Transaction | Booking completes in English | Booking form loads in English | 0 |
| Freshness | Updated four months ago | Updated twenty-two months ago | 0 |
| Review | Clinically reviewed | Translated after review, not re-reviewed | 1 |
Result. 5 of 16.
Reading it. Coverage parity is 2, which is why the organization believes it has a bilingual site. Every dimension that affects whether a patient can act scores 0 or 1. This is the characteristic profile of a translated site: present everywhere, decisive nowhere.
Decision implication. The cheapest gains are practitioner parity and access parity, because the information already exists in the English version and needs placing, not creating. Transaction parity is the most expensive and the most valuable, since it is the point of loss. Discovery parity requires new research and should be scheduled rather than rushed.
Limitations. Scoring is judgement-based against defined criteria. The composite is constructed to show the pattern, not to represent a measured average.
08
Governance for bilingual publishing
Parity is maintained by workflow, not by intention. Four mechanisms hold it.
Paired publishing states. The content system records the relationship: both published, Arabic pending, English updated and Arabic review required, or intentionally non-equivalent with a documented reason. Without these states, drift is invisible.
Arabic-native review. Clinical and compliance review of Arabic content must be conducted in Arabic by a qualified reviewer. Reviewing the English and approving the translation is not review, since the errors that matter appear in the Arabic.
Shared update triggers. A consultant joining, a service moving, an insurance change or a regulatory update triggers both language versions in the same task, not two tasks with different owners and different priorities.
Independent measurement. Arabic performance is reported separately. Aggregating both languages into one traffic figure conceals precisely the gap this framework exists to expose.
The broader operating model for this sits in Medical Content Governance: Ownership, Review, Updates and Retirement.
09
Priorities
Fix now
- Bring practitioner profiles to parity, using detail that already exists in English
- State insurance acceptance, hours and site availability on Arabic pages
- Fix right-to-left rendering errors in numbers, phone numbers and Latin terms
- Confirm hreflang and self-referencing canonicals are correctly implemented on every pair
Build next
- Conduct Arabic keyword research natively, from Arabic query data, for the priority service lines
- Bring the booking path to full Arabic parity including confirmation messages
- Rewrite Arabic service pages to equivalent depth rather than editing translations
- Establish paired publishing states and Arabic-native clinical review
Measure continuously
- Parity score by page pair for priority service lines
- Arabic non-branded entries reported separately from English
- Arabic booking completion rate against English
- Freshness gap in days between paired updates
10
Benchmark methodology and limitations
What this article is. A framework for measuring bilingual parity in healthcare, with a scoring instrument. It is an editorial and analytical contribution built from platform guidance and recurring patterns in Saudi healthcare websites.
What it is not. The page-pair comparison is a constructed composite. No DEMA bilingual parity benchmark has been published, so no average parity score for Saudi healthcare sites is offered.
Method limitations. Parity scoring involves judgement and should be applied consistently by the same method. Depth parity in particular requires assessment of decision value rather than word count, since equivalent meaning does not require equal length.
Where review is required. Arabic clinical content requires review by a qualified reviewer working in Arabic. Regulatory claims require verification against current Saudi requirements in the language of publication. Booking changes require privacy review under the Personal Data Protection Law. This article is not legal advice.
Planned research. DEMA has scoped a Saudi bilingual trust parity study. Until it is completed, no empirical parity findings exist and none should be inferred here.
Sources.
11
Does your Arabic site sell as well as your English one?
Almost nobody knows, because the two are rarely measured apart. A Healthcare Growth Diagnosis scores your priority page pairs across all eight parity dimensions and shows where Arabic-speaking patients are being asked to decide with less.
Request a Healthcare Growth Diagnosis