Introduction

A person with two or more long-term conditions may spend a considerable part of ordinary life arranging appointments, collecting prescriptions, monitoring symptoms, changing meals, interpreting advice from different clinicians and paying for transport or medicines. These tasks are not the same as symptoms caused by disease. Nor is their difficulty captured by a medication count or a record of missed doses. The cumulative-complexity model describes burden as an imbalance between the work that care and life demand and the capacity available to do it; minimally disruptive medicine asks clinicians to reduce unnecessary work while pursuing outcomes that matter to the patient. [1, 2] A measure of treatment burden should therefore tell us something about experienced work and impact, rather than relabel adherence or disease severity.

Questionnaires can help a busy service ask about this work consistently. They can also mislead. A patient who reports little burden may have adequate support, may have stopped a demanding regimen, may interpret a translated item differently, or may be reluctant to report difficulty to a clinician. A high score may arise from a particular scheduling problem that a single total cannot identify. The 2023 review of burden-of-treatment measures in multimorbidity found eight instruments across 72 studies but described gaps in content validity, responsiveness, floor effects and score interpretation. [3] An integrative review mapped the work and impact of multimorbidity across 30 studies and noted that commonly used instruments missed the continual prioritisation that patients do when tasks compete. [4] A later systematic review of PROMs in chronic comorbidities again found sparse cross-cultural and responsiveness evidence. [5] A 2026 narrative review addressed the Treatment Burden Questionnaire in geriatric care. [6] The incremental contribution here is an intended-use decision framework for Saudi primary care, grounded in Arabic adaptations and the ambiguities of the Jeddah study, rather than another inventory. The added focus is a critical comparison of instrument purpose, intended use and transferability, not a claim that Arabic treatment-burden tools are absent.

There is already Arabic-language work on this topic. The Multimorbidity Treatment Burden Questionnaire has an Arabic adaptation, MTBQ-A, tested among adults with multimorbidity in the United Arab Emirates. [7] Arabic versions of a medicines-specific quality-of-life measure have progressed from content adaptation through exploratory and confirmatory testing to a hospital feasibility study. [8, 9, 10, 11] An Arabic translation of the Living with Medicines Questionnaire predates those reports. [12] A Saudi primary-care study has also reported MTBQ scores in Jeddah. [13] Consequently, a claim that Arabic measures do not exist, or that a new translation is the obvious first step, would be wrong.

The relevant Saudi question is more demanding: When a primary-care clinic wants to learn how much work its treatment imposes on Arabic-speaking adults with multimorbidity, which measure fits the decision, which properties have been tested in comparable people, and what evidence is missing before scores guide triage or service evaluation? This review compares constructs and settings rather than ranking instruments with a formal psychometric score. It proposes a bounded evaluation sequence, not a new scale or an implemented Saudi screening programme.

Sources and scope

Three orientation searches were run through PubMed E-utilities at 16:06 UTC on 27 September 2026, without date or language limits. The regional query was ("treatment burden"[tiab] OR "burden of treatment"[tiab]) AND (Arabic[tiab] OR Saudi[tiab]) AND (multimorbid[tiab] OR medication[tiab]) (seven returns). The instrument query was ("Multimorbidity Treatment Burden Questionnaire"[tiab] OR "Medication-Related Burden Quality of Life"[tiab] OR "Patient Experience with Treatment and Self-management"[tiab] OR "Treatment Burden Questionnaire"[tiab]) AND (validat[tiab] OR psychometr[tiab] OR translation[tiab]) (34 returns). The review query was ("treatment burden"[tiab] OR "burden of treatment"[tiab]) AND (measure[tiab] OR instrument[tiab]) AND (review[pt] OR "systematic review"[tiab] OR "scoping review"[tiab]) (146 returns). The exact returned identifier lists and request URLs are retained in a source log. Returns overlap; they were not all screened, deduplicated or assessed by two reviewers. SciSpace semantic search, title/DOI lookup and references of key papers were used to locate Arabic and international antecedents that a single PubMed phrase might miss. The SciSpace results were discovery leads, not evidence until source identity and relevant claims were checked in PubMed, Crossref or an accessible primary article.

Selection was purposive: we retained primary development/adaptation studies for instruments measuring treatment work or medicines' impact; Arabic versions or applications relevant to Gulf care; the Saudi Jeddah use study; and earlier reviews that define overlap. We excluded disease symptoms, satisfaction and patient activation as burden measures, although adjacent constructs are discussed where confusion matters. No number of excluded papers is asserted because the overlapping search returns were not completely screened. The full texts of the Jeddah study, original MTBQ work, several Arabic MRB-QoL reports and selected reviews were examined through PubMed Central/Europe PMC. MTBQ-A and the 2026 Jazan qualitative report were checked at abstract level; claims do not extend beyond their accessible records. Targeted web-index checks on 27 September 2026 using site:imemr.net and site:search.emarefa.net with "treatment burden" Arabic review found no directly matching Arabic-instrument review among returned results; this is not a database-wide negative search. We did not request unavailable publisher articles, perform paired screening, formally rate risk of bias or measurement quality, calculate a pooled estimate, or claim to have found every Arabic instrument. The counts above are search returns, not a systematic-review flow. This article is a critical narrative review, not a scoping or systematic review. Table 1 makes the central source access and selection boundary visible.

Table 1. Central source families selected for this narrative decision question

Table 1 footnote: This purposive map is neither an included-study denominator nor a formal COSMIN appraisal. “Access” describes what was available during this review, not all material that a journal or institution could obtain. The accompanying author research archive retains exact PMID lists, retrieval timestamp and source-level access notes; release of those records as a supplement requires accountable author approval.

What a treatment-burden score should represent

Treatment workload has multiple components. A patient must understand what to do, make the regimen compatible with work and family, reach services, manage medicines and monitoring, and repair conflicting plans. The consequences include fatigue, lost time, financial pressure, reduced social participation and strain in relationships with clinicians. [1, 19, 4] Some instruments ask mainly about the difficulty of doing tasks; others ask about how medicines affect functioning and well-being. Both can be useful, but they answer different questions. A Saudi clinic evaluating appointment coordination may learn little from a medicine-only impact score. A pharmacist reviewing a complex regimen may find that same score more actionable than a broad total.

Adherence is related yet distinct. A patient can follow a burdensome plan at great personal cost, or can have low recorded adherence because the plan is unaffordable, confusing, clinically inappropriate or simply not aligned with the patient's goals. The original Treatment Burden Questionnaire was negatively correlated with treatment satisfaction, and the English adaptation found higher burden among people reporting lower medication adherence; those associations did not make the scales interchangeable. [17, 18] A Saudi Arabic treatment-satisfaction validation in warfarin clinics measures satisfaction with anticoagulant therapy, not the full work of living with several conditions. [20] The Arabic Patient Activation Measure tested in Riyadh primary care assesses knowledge, beliefs, confidence and skills, which are relevant to capacity, but it is not a direct treatment-workload measure. [21] Choosing the familiar questionnaire because it is already translated would answer the wrong question if the intended outcome is burden.

The unit of interpretation also matters. A one-time score can start a discussion, while a change score requires evidence that the scale can detect meaningful change beyond measurement error in that setting. Comparison across education, age, nationality or clinic types requires evidence that the items function similarly across those groups. A high internal-consistency coefficient does not establish measurement invariance or responsiveness: it tells us that items covary in one sample, and may even rise when items are redundant. If a service intends to use a cutoff to trigger extra appointments, its sensitivity, false positives, missing-data behavior and consequences need testing under the intended workflow. The evidence for simply eliciting an item-level concern is less demanding than that for labelling a patient “high burden” or comparing hospitals.

Administration is part of what a score means. An interviewer can help someone who has difficulty reading, but a patient may then hesitate to disclose cost, confusion or dissatisfaction, particularly if the interviewer is a member of the treating team. A self-completed digital form reduces that interpersonal pressure for some people while excluding others who cannot access it or who need assistance. Proxy completion by a family member may capture real family labour yet should not be silently merged with the patient's own report. A Saudi validation should record mode, language variety, assistance and relationship to the clinic for every response and examine whether item distributions change across these conditions. The approach must preserve the patient's ability to describe a burden that the fixed response options did not anticipate. These are measurement-design requirements inferred from the construct and the broader Arabic PROM literature, not outcomes demonstrated by the existing Saudi MTBQ study. [22, 23]

Established instruments and the Arabic evidence

The 10-item MTBQ was developed in UK primary care using literature, patient involvement and cognitive interviews, then tested in 1,546 adults with multimorbidity whose mean age was 71. It supported a one-factor structure and Cronbach alpha of 0.83; the authors observed expected correlations with other burden measures and quality of life and examined change over nine months. [14] This gives the original score a coherent development history, not automatic validity in Arabic-speaking or younger Saudi populations. Further UK testing in 244 adults found test–retest agreement of 0.765, yet reversing the response scale did not solve floor effects or skewness. [24] A score crowded near zero can miss change among people whose difficulties are concentrated in one neglected domain.

The MTBQ-A is an evaluated Arabic-language adaptation tested among UAE community-pharmacy patients, not a validation study of Saudi primary-care patients. Professional forward–backward translation, expert review, cognitive interviews and pilot work preceded testing in 177 adults with multimorbidity recruited through UAE community pharmacies. In the 177-participant UAE validation, the reported content-validity index was 0.94 and internal-consistency coefficient was 0.83; observed correlations supported construct validity within that source setting but not an established Saudi clinical threshold. Factor analysis supported two dimensions, described by the authors as self-management/social support and burden of visiting services and professionals, rather than the one-factor pattern of the original UK study. Response distributions were positively skewed with high floor effects. [7] These findings warrant use in the population studied; they do not establish that the UK global thresholds, factor scoring or sensitivity to intervention change transfer unchanged to Saudi primary care. Nor does the abstract provide a Saudi dialect or literacy analysis.

A floor effect has practical consequences beyond a psychometric label. If many people select the lowest response across most items, a later reduction in one concrete difficulty may leave the total unchanged. Conversely, a single item about service access can worsen while the total remains low. The two-factor finding also means that a clinic should not assume a single sum captures the same construct as the original UK score without testing that interpretation. Even if a locally fitted model resembles the UAE model, this would not by itself establish a clinically useful cutoff or a patient-important change. Those require anchors linked to what patients say has improved, with an explicit time frame and adequate follow-up. The later UK study's persistence of floor effects after response-scale reversal suggests that merely changing the direction of response options is unlikely to solve the underlying content or distribution problem. [24]

Broader comparisons are useful precisely because factor structure can vary across translations and settings. Danish, German and Dutch adaptations of the MTBQ each studied their own populations and measurement properties; their existence illustrates why a translated name is not a universal psychometric result. [25, 26, 27] The question for a Saudi service is whether the tasks named by the items, response options and time frame fit its own care system and patient lives. An item about travel or obtaining appointments may behave differently when care is dispersed across providers, when a family member arranges visits, or when medicines are supplied through several systems. These are hypotheses for local interviews and analysis, not findings already established by the Gulf validation.

The 31-item Medication-Related Burden Quality of Life tool, MRB-QoL, began in Australia with an emphasis on the impact of medicines on daily functioning and well-being. Its initial study in 367 respondents produced five domains, including regimen complexity, psychological and social burden, functional roles and therapeutic relationships. [28] Arabic adaptation involved expert review and patient cognitive interviews and reported a final content-validity index of 0.96. That 2024 report correctly asked for additional psychometric work before broad clinical use. [8] Subsequent UAE studies supplied some of that work: exploratory testing among 400 hospital patients found a four-factor Arabic structure, alpha 0.973, test–retest intraclass correlation 0.994 and a reported minimum detectable change of 3.89 points under that study's conditions. [9] Confirmatory testing in 343 adults from four UAE tertiary hospitals supported first- and second-order models of the four-domain Arabic version; its factor structure differs from the original English five-domain pattern. [10] These are important advances, but very high reliability coefficients are not evidence that every patient will find the questions relevant or that a four-point change is a meaningful improvement in Saudi primary care.

A mixed-methods UAE study then had 10 clinical pharmacists use MRB-QoL Arabic during routine medication reviews for 227 hospital patients, with a mean System Usability Scale score of 82.2. [11] Its non-randomised, non-controlled design provides preliminary feasibility and acceptability evidence, not an estimate of better adherence, fewer admissions or reduced patient burden caused by the questionnaire. The hospital population and pharmacist-led workflow are also different from a general Saudi primary-care consultation. The appropriate next question is whether the information changes a jointly agreed plan in a way patients recognize as helpful, without adding more work.

An earlier Arabic Living with Medicines Questionnaire translation addressed medicines-related burden through forward/back translation and cognitive debriefing, with developer permission; its 2016 report called for further validation. [12] The 15-item Treatment Burden Questionnaire was also administered in Arabic to 280 patients with advanced chronic kidney disease in Qatar; a subsequent paper from that cohort describes linguistic translation, but a disease-specific application does not establish broad Saudi multimorbidity validity or a primary-care cutoff. [15, 16] The original French and English TBQ and the more extensive Patient Experience with Treatment and Self-management scale, PETS, bring broader task and impact domains but cannot be described as Saudi-validated Arabic alternatives. The French TBQ was tested in 502 patients, and its English adaptation in 610; PETS development and later work provide domain, test–retest and change evidence in US samples. [17, 18, 19, 29, 30] A 2026 short-questionnaire development report used cognitive interviews with 15 UK adults; it does not itself validate a Saudi Arabic short form. [31] Table 2 places these families against the decision they could support without suggesting that each is ready for local threshold use.

Table 2. Selected treatment-burden instruments and the limits of transfer to Saudi primary care

Table 2 footnote: The rows distinguish instrument purpose and source populations; they are not a formal COSMIN ranking or an exhaustive register of Arabic tools. “No comparable Saudi validation found” describes targeted searches, not proof of absence. Factor counts, coefficients and sample sizes refer to cited studies and should not be pooled. Use and translation rights must be checked with each instrument owner before local deployment.

What the Saudi observations do and do not establish

A Jeddah cross-sectional study recruited 422 people with at least two clinically diagnosed long-term conditions from 12 primary-care centres in late 2023. It administered the MTBQ alongside a medication-adherence scale and labelled 232/422 (55%) “high burden”, with lower adherence scores in that category. [13] This is the source's label, not a national prevalence estimate, intervention effect or validated Saudi triage rule. The accessible full paper cites the original UK MTBQ but does not specify a numeric high/medium/low scoring rule, the exact Arabic wording administered, use or equivalence of the separate UAE MTBQ-A, or a local cognitive-testing route. Its stated eligibility is age 18 years or older while Table 1 reports an age range of 15–88; its occupation significance statement also conflicts with the printed table P value. It excluded several groups, including patients in palliative care, active cancer and selected mental health diagnoses, and did not perform multivariable regression because assumptions failed in small categories. These inconsistencies do not negate the participants' reported difficulties, but require source-specific caution and preclude transferring its category or prevalence estimate to other clinics.

The numerical category is particularly sensitive to the scoring rule. The original MTBQ investigators grouped nonzero scores into sample tertiles for interpretation; this is a research convention, not a universal biological boundary. [14] A later clinic that imports “high” as a fixed clinical label without checking the Jeddah scoring procedure, version, translation and case mix could classify people differently for reasons unrelated to their experience. The Jeddah study reports a cross-sectional association between burden and adherence, but does not determine whether demanding treatment reduces adherence, whether patients struggling with adherence perceive greater burden, or whether social circumstances influence both. The fact that active cancer and some mental health conditions were excluded further narrows transfer to a general multimorbidity clinic. A Saudi service should therefore retain raw item and domain patterns alongside any exploratory category and state exactly who was eligible to complete the form.

The 2026 Jazan qualitative report interviewed 19 adults taking at least five long-term medicines and described information work, self-devised medication routines, psychological strain and the relevance of Ramadan. [32] Only the indexed abstract was available for this review. It helps frame possible interview prompts and contextual hypotheses; it cannot validate a questionnaire, quantify regional burden or prove that a particular item has different meaning during Ramadan. Local patient participation must determine whether and how such topics belong in the measure or clinical conversation. A medicine count by itself remains a poor substitute for that account of effort and capacity.

These observations also expose a practical tension. A broad scale may make service access visible while overlooking a patient whose main difficulty is the adverse effect or cost of one medicine. A medicine-focused scale can uncover that problem yet leave appointment travel, conflicting advice and self-monitoring outside its score. A clinic may need one brief general opening question and a more detailed domain probe only when relevant. This is a design hypothesis, not a validated combination of instruments. Repeated questionnaire completion itself is treatment work; the service should show that it acts on responses before asking patients to supply more data. The 2022 and 2023 reviews of Arabic PROMs reported that cross-cultural validity, responsiveness and measurement error have often been underexamined across fields, reinforcing the need to test the intended use, not merely translate more items. [22, 23]

From score to conversation, evaluation and governance

For an individual consultation, a clinician may ask open, plain-language questions about which care task or effect is most difficult, invite correction and agree whether a safe change in the care plan is possible. This does not require administering, copying or scoring a proprietary questionnaire. Using a published item's actual wording, even informally, requires a version and permission check; a reproducible score also requires local meaning and workflow testing. If a locally authorized instrument is later used, a low total should not end the conversation when a single item signals difficulty, and a high total should not automatically trigger extra appointments or medicine changes. The patient may want fewer visits, a clearer plan, home delivery, a different dosing schedule or help coordinating specialists; each action has its own clinical safety context. A questionnaire is not itself a diagnosis, an adherence judgment or a treatment order. [1, 2, 19]

For service evaluation, the threshold is higher. If a clinic uses MTBQ-A to compare sites or dates, it needs a defined eligible population, consistent administration mode and language, recorded nonresponse, and evidence that scores mean comparable things across groups. It must also prespecify case-mix variables and describe who can and cannot complete the form. Differences may reflect disease severity and mix, selection into care or survey response, administration mode and the services available, rather than clinic quality. If a pharmacist uses MRB-QoL Arabic before and after a medicine review, measurement error and responsiveness must be tested in that Saudi setting; its UAE minimum detectable change of 3.89 is a study-specific error estimate, not a patient-important improvement threshold. [9] The output should sit beside balancing outcomes: medicine safety, visits, cost, time spent completing questionnaires and patient-reported benefit. Collecting forms while leaving care unchanged may increase the burden being measured.

Table 3 states a proposed evaluation sequence. The COSMIN PROM framework distinguishes content and structural validity, reliability, measurement error, responsiveness and interpretability; this review uses those domains to explain why intended uses demand different evidence but does not conduct a COSMIN review or rating. [33] The local interview, workflow, case-mix and safety steps in Table 3 are the review authors' proposals, not published COSMIN requirements, nor a checklist endorsed by MHJ or a Saudi clinical society. The sequence should be designed with Arabic-speaking patients of varying education, age, disease combinations and care arrangements, including those who need help reading. Researchers should specify whether the goal is a conversation, group comparison, change detection or triage before choosing an instrument. That choice determines the indispensable measurement evidence and whether an extensive scale is justified in a short consultation.

Table 3. Proposed evidence needed for four intended uses in Saudi primary care

Table 3 footnote: Content/structural validity, reliability, error and responsiveness are recognized measurement domains in COSMIN guidance [33]; the specified Saudi workflow, case-mix, safety and clinical-impact steps are the review authors' proposals. This table supplies no validated Saudi thresholds, individualized rules or formal COSMIN rating. Clinical, information-governance and patient representatives must approve any implemented workflow and confirm instrument-use terms.

The study design should separate measurement testing from clinical-effect testing. First, qualitative interviews can reveal concepts the selected instrument misses and items whose interpretation depends on clinic or family context. A small pilot can then measure completion burden and missingness. A larger prospective validation can test structure, reliability and cross-group comparability; a repeated-measures cohort with a prespecified anchor can examine change. Only after the score has a credible interpretation should a controlled or carefully designed pragmatic evaluation ask whether feeding results back to clinicians improves a patient-important outcome. Training, time to review scores and a route to act on problems are part of the intervention, not background assumptions. The UAE pharmacist feasibility study demonstrates one way to examine usability, but it neither substitutes for Saudi transport testing nor proves benefit. [11]

For that final evaluation, the comparison should be ordinary care with the same opportunity for patients to raise concerns, not a group denied necessary clinical attention. Teams could predefine actions after an item response—for example, a medicine-review referral for a regimen problem or a scheduling review for repeated visits—while leaving individualized decisions to the treating clinicians. Outcomes should include the patient's own account of whether work became more manageable, completed changes to the care plan, safety events, unplanned care and the time patients and staff spent on the new process. A process measure such as “forms completed” is insufficient: it can improve even as unresolved burden grows. Missing forms matter too, because the people with least capacity may be least likely to complete them. Interpretation should prespecify how those missing responses are described and avoid treating them as zero burden. The goal of this sequence is to determine whether measurement changes care in a useful and safe direction, rather than to reward a high response rate alone.

Rights and accountability are part of the evaluation. The Living with Medicines Questionnaire translation reported obtaining developer permission; other instruments may have different access and reproduction terms. [12] Any clinic intending to copy items into an electronic record or translate them further must verify the current instrument version and use terms. Data collection also requires a responsible clinical team, a plan for responding to urgent safety concerns, and appropriate institutional information-governance decisions. This review uses published studies and no new patient data; it cannot supply approvals or consent for a future validation study.

Interpretation and limitations

This review makes a narrower claim than a psychometric synthesis. It does not formally rate the quality of each study, establish measurement invariance, recommend one instrument for every setting, or determine a Saudi score threshold. The source search was targeted and may have missed non-indexed Arabic work. Some pivotal records, including MTBQ-A and the Jazan interview study, were assessed from abstracts; the UAE MRB-QoL and Jeddah full texts were examined more closely. This uneven access limits fine-grained judgments about sampling, missing data and item performance. The broad systematic and scoping reviews cited here have their own populations and time windows; their counts cannot be treated as a fresh search completed for this article. [3, 22, 23, 5]

The contrast between MTBQ-A and MRB-QoL Arabic is primarily a contrast of construct and setting, not a head-to-head performance comparison. No identified primary study administered both in the same Saudi primary-care sample. The Jeddah study's “high burden” category is its own reported category, not evidence that the UK or UAE cutoff is ready for universal Saudi triage. The strongest near-term conclusion is that there are Arabic tools worth evaluating, while the strongest open question is how they behave and affect care when used in the intended Saudi service. No identified Saudi primary-care study directly compared these instruments or established that applying their scores improves clinical decision-making.

Conclusion

Arabic treatment-burden measurement should begin with the decision it is meant to inform. MTBQ-A offers a brief view of multimorbidity care work; MRB-QoL Arabic explores medicines' impact in greater detail; Arabic LMQ and TBQ applications provide other antecedents with different limits. A Saudi clinician can ask open questions in plain language about what care costs a patient in time, effort and opportunity. A clinic should administer and score a named questionnaire only after checking its current version, permission and local meaning; it should not use an imported cutoff as a diagnostic label or quality target until comparability, change interpretation and clinical consequences have been tested. A patient-centred validation and implementation study is a more defensible next step than another uncontextualized prevalence percentage.