Health Literacy and Health Outcomes in Patients with Back Pain – A Scoping Review

BACKGROUND: The World Health Organisation highlights back pain as a leading and growing cause of disability worldwide. In the eld of chronic pain, higher health literacy is linked to lower pain intensity and better pain control. However, there remains a paucity of evidence, with a recent systematic review nding only three studies meeting its inclusion criteria. OBJECTIVES: This scoping review had two objectives in exploring research in chronic pain, which were to identify: 1) The back-pain health outcomes included in such work and the extent to which these reect the core outcome set for clinical trials in non-specic low back pain. METHODS: The search broadened the search strategy used in the systematic review. It was conducted using thirteen bibliographic databases, employing medical subject heading (MeSH) terms for back pain and health literacy, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines. A thematic framework approach was used for analysis. RESULTS: The search yielded ten relevant studies for inclusion, amongst which a total of nine health literacy measures and 50 health outcome measures were used. The health outcomes assessed by the included studies could be broadly categorised into: Pain, Disability, Behaviour, Knowledge and Beliefs, and Resource Utilisation. Most outcome measures employed (36 out of 50) were not directly relevant to addressing the core outcome set for clinical trials in non-specic low back pain. CONCLUSIONS: A dearth of studies in this eld of research was noted, especially in Asia and low-middle income countries. Methodological limitations were identied, including in study design and outcome measures. To allow for comparison across ndings and the development of a rigorous evidence base, future work should include the core outcome set for clinical trials in non-specic low back pain. Furthermore, research thus far has focused on a narrow range of populations and there is an urgent need to broaden the evidence-base to include those with characteristics known to be moderators in health outcomes. Such work demands the incorporation of comprehensive measures of health literacy that have generic and culturally sensitive components. Provide a structured summary including, as applicable: background, objectives, eligibility criteria, sources of evidence, charting methods, results and conclusions that relate to the review question(s) and objective(s).


Background
The concept of health literacy (HL) is extensive and incorporates functional, communicative/interactive and critical domains (1). It is de ned as "the achievement of a level of knowledge, personal skills and con dence to take action to improve personal and community health by changing personal lifestyles and living conditions" (2). At its core is an observable set of skills that can be developed and improved through effective communication and education to enhance autonomy and empower people to make decisions relating to their health and changing circumstances (2,3). At the inaugural Outcome Measures in Rheumatology Clinical Trials (OMERACT) Health Literacy Special Interest Group workshop, 16 themes at the micro, meso and macro level were identi ed, including cognitive capacity, access to information, and health systems (4). Independent of other socio-demographic factors low HL is associated with higher mortality amongst older people, with poorer health outcomes, and with higher morbidity more generally (5)(6)(7), whilst higher HL is linked to lower pain intensity and better pain control among those with chronic pain (8,9).
Back pain is the single leading cause of disability globally and is rising (9,10). Financial costs from back pain are estimated to be in the order of billions of US dollars (USD) (10,11), while the economic burden of members of the workforce suffering from back pain is estimated in the USA alone to be USD 7.4 billion/year (12). Traditionally conceptualised as solely secondary to mechanical injury, back pain is now described within a bio-psychosocial model, resulting from an interaction of physical, psychological and social in uences (13). Effective self-management is crucial to improving back pain outcomes (14,15). Studies have also demonstrated the need to focus on HL in order to develop effective patient education materials and/or patient resources to support self-management in such patients (16,17).
However, inconsistent reporting of outcomes in clinical trials of patients with non-speci c low back pain has been highlighted, together with the implications for comparing ndings and the reliability of systematic reviews. To address this issue a core outcome set (COS) has been developed, led by an International Steering Committee, de ning the minimum set of outcomes that should be reported in all clinical trials. The COS includes 'physical functioning', 'pain intensity', 'health-related quality of life' and 'number of deaths' (18).
Asia has the largest number of low back pain disability-adjusted life years internationally and the highest risk of occupational back pain is in the agricultural sector -a major sector in Asian economies (19). In addition, there exists a paucity of research to underpin evidence-based practice of low back pain treatment in low-and middle-income countries (LMIC) (20,21). Existing evidence tends to be from developed Western nations and cannot be accurately applied to the Asian and LMIC context, given that pain reporting, manifestation and management is in uenced by socio-cultural and genetic factors (22).
To develop more evidence-based interventions and guidelines we need to better understand the relationship between HL and back pain outcomes. A previous systematic review by Edward et al. in 2018 into the impact of HL on low back pain management identi ed only three relevant studies, all of which were based in Western nations. However, the review acknowledged "possible incomplete retrieval of identi ed research and reporting bias" (23) as the search was limited to four bibliographic databases and limits were also placed on year of publication, language, and article formats, amongst other search lters (23).
This scoping review builds on Edward et al.'s work and had two objectives. These were to identify: 1. The health literacy measures currently employed and the aspects of health literacy they include (Functional, Communicative and Critical). 2. The back-pain health outcomes included in such work and the extent to which these re ect the core outcome set for clinical trials in non-speci c low back pain.
To mitigate the shortcomings of incomplete data retrieval which limited Edward et al.'s work, this study expanded the search from four to 13 bibliographic databases and did not utilise search limiters or lters such as time or language lters. Unlike the systematic review carried out by Edward et al., this study is a scoping review with the emphasis not on reporting the degree of association between HL and back pain health outcomes, but on scoping the variety of HL and health outcome measures employed in existing literature, while additionally providing a critique on the choice of outcomes studied and measures used in order to identify implications for future research.

Inclusion and Exclusion criteria
The Joanna Briggs Institute (JBI) manual's PCC mnemonic (24) was used to clarify the research focus: Population -Patients with back pain (≥ 10% of study population), of any age, gender, or race Concept -Relationship of back pain health outcomes to HL Context -Original primary study (any study design), in any healthcare setting In addition, studies for inclusion required the use of speci c HL and health outcome measures. Studies were excluded if they only analysed generic literacy, numeracy, and education level not in the context of healthcare. Generic patient education interventions have the potential to in uence non-HL related determinants of back pain, hence drawing conclusions about HL's effects on back pain from these studies may be inaccurate (25), and for this reason these studies were excluded.

Study selection, data extraction and analysis
The search strategy followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines (26). Independent review of titles and abstracts from the initial search was conducted by two reviewers (CS and WW). Studies then underwent a full-text review if they investigated a relationship between HL and back pain outcomes. A third reviewer (HES) resolved any discordance between reviewers.
Data extraction included determinants of HL (age, gender, ethnicity, and education level) (27), study design, and measures and types of outcomes used. Adopting a framework analysis approach (28,29), the health outcomes used were collated and coded into descriptive themes, and then grouped into overarching categories. Discrepancies were resolved through discussion between reviewers. These categories were then mapped against the core outcome domains for clinical trials in non-speci c low back pain (18), namely 'physical functioning', 'pain intensity', 'health-related quality of life' and 'number of deaths'. The HL measures used were also categorised according to the components of HL they covered, using the classi cation proposed by Nutbeam, namely Functional, Communicative and Critical (30,31). Functional HL encompasses basic skills in reading and writing, Communicative HL includes social skills and advanced cognitive and literacy skills to actively participate in daily activities, and Critical HL comprises the use of even more advanced cognitive and social skills to exert great control over life events and situations.
As this was a scoping review grading of evidence was not conducted. Instead, this study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines (26), as detailed in Appendix 2. In line with best scoping review practice (32), two experts, one in the eld of HL (JP) and one in back pain (JH) were consulted to further inform the study's ndings.

Articles reviewed
The initial search yielded 5508 articles. After removing duplicates and reviewing titles and abstracts, 17 articles remained for full-text review. Ten of these were included in the nal analysis (Fig. 1). Since each article reported a different HL measure, both 2010 and 2011 papers by Briggs et al. (15,33) were included. Table 1 here. 3.2 Summary of key data retrieved from full-text reviews Despite no restrictions being placed on year of publication, all the studies meeting the inclusion criteria were published in 2010 or later, and all were cross-sectional in design (Table 1). They employed structured questionnaires, apart from one mixed methods study which also used interviews. In terms of country of origin, two studies were conducted in Australia, four in the United States of America, three in Europe (Germany, Austria, Switzerland and Italy), one in Saudi Arabia. Five of the cross-sectional studies solely studied patients with back pain, while the other ve studies included back pain as a signi cant portion of their patient population (> 10%), one of which provided a subpopulation analysis of patients with back pain.
All ten studies focused on adult populations (≥ 18 years) and had a majority female population, with one study having entirely female participants. Ethnic break-down was only provided by two USA studies, both of which had predominantly white study populations. Nine studies collected data on education level, most reporting an even spread across participants. Table 2 here. Beliefs on "cause and course of low back pain", and "seeking, understanding and utilising low back pain information".

Telephone interviews
Unspeci ed correlation to HL. However, participants reported obstacles in seeking, comprehending and using low back pain information, which were not re ected in S-TOFHLA scores. Utilisation of lumbar spine treatment (last 6 months), physiotherapy attendance, immunisation history, medication use, employment, days of work missed.
Adequate HL group used more medications and consulted a specialist more frequently than limited HL group.
Limited HL group reported more individual visits to chiropractor and had lower use of NSAIDs. Healthcare Access Healthcare provider visits (last 12 months), health insurance coverage, delayed healthcare due to cost concerns, delayed healthcare due to non-cost concerns, ability to afford common supplementary healthcare.
Low HL associated with greater inability to afford ancillary care.

Summary of HL measures used
Nine different HL measures were employed across the ten studies (Table 2). Most HL measures assessed functional HL (72.7%), while the number that evaluated communicative and critical HL were fewer than half (27.3% and 36.4% respectively) ( Table 3).
In their 2010 paper, Briggs et al. (15) used the short-form Test of Functional Health Literacy in Adults (S-TOFHLA), which assesses 'functional' HL, de ned as reading, writing and numeracy skills in relation to healthcare (1,34). The S-TOFHLA does so via the use of two prose passages and four items testing numeracy. Briggs et al. (15) also employed telephone interviews to assess HL by asking participants on how they sought, understood and utilised back pain information.
Subsequently in their 2011 paper, Briggs et al. used the Health Literacy Measurement Scale (HeLMS) (33), which goes beyond functional HL to include communication skills, computation skills, and social support, thereby overcoming limitations of the S-TOFHLA (33). The HeLMS sets out to assess "overall capacity to seek, understand and use health information within the healthcare setting" by asking questions such as "Are you able to see a doctor when you need to?" (33). By doing so it assesses all three domains of health literacy, Functional, Communicative and Critical HL.
The Newest Vital Sign (NVS) was the most frequently used HL measure (Table 3), utilised by Glassman et al., Al-Eisa, Buragadda and Melam, and Devraj, Herndon and Gri n (17,35,36). The NVS is convenient to use and has a sensitivity equivalent to the TOFHLA for identifying inadequate HL. It assesses functional HL through the asking of six questions regarding a standardised ice cream nutrition label (37).
Besides the NVS, Glassman et al. also used The Health Literacy Assessment, a 10-item self-administered questionnaire using items selected from the computerized Health LiTT measure (35). The Health Literacy Assessment (Health LiTT) assesses HL via three sections: Prose, Document and Quantitative (38). The Prose section asks participants to ll in missing words in a cloze passage, while the Document section consists of multiple-choice questions regarding images such as a prescription label. The Quantitative section also uses multiple-choice questions requiring arithmetic computation. By not assessing social skills and life choice decision making, the Health Literacy Assessment (Health LiTT) con nes itself to Functional HL assessment only.
Camerini and Schulz (39) interpreted HL based on scores from the Low Back Pain Knowledge Questionnaire. The Questionnaire involved multiple-choice questions on topics such as the aetiology and management of low back pain. This neither quanti es one's reading, numeracy nor social skills, but arguably involves decision making over treatment options and can be categorised as assessing critical HL.
Measurement of critical HL alone is also seen in other studies. MacLeod et al. used a single-item screener "How con dent are you lling out medical forms by yourself?" (40), while Burke, Nahin and Stussman used the response option "Never heard of it/Do not know much about it" from the National Health Interview Survey (41).

Köppen et al. used HL questions taken from the Brief Questions to Identify Patients with Inadequate Health Literacy (8).
These included the questions "how often do you have someone help you read hospital materials", "how con dent are you lling out medical forms by yourself" and "how often do you have problems learning about your medical condition because of di culty understanding written information?" (8). These questions do offer a brief assessment of one's reading and social skills, hence evaluating functional and communicative HL.
Finally, Farin, Ullrich and Nagl developed the HELP questionnaire (Health Education Literacy of Patients with chronic musculoskeletal diseases), an 18-item assessment that aims to summarise a patient's reported communication and comprehension di culties in health education and treatment (16). Questions such as "How much di culty did you have communicating your own expectations and wishes in terms of your therapy?" were scored on Likert scales anchored from 1 to 5, with lower values indicating a higher HL. This instrument does not directly evaluate control over life events and choices and is categorised as covering only functional and communicative HL.    (Table 4).

Pain intensity was also quanti ed as a sub-component of the Chronic Pain Grading (39), the Short-form McGill Pain
Questionnaire (8), the Oswestry Disability Index (ODI) (15,33,35,36), and the Euro-QOL5D (35). In addition, the Shortform McGill Pain Questionnaire assesses the nature of pain (42), while other studies looked at pain duration and frequency (Table 4).

Disability
Five studies involved data on disability (15,33,35,36,39). Lastly, the Chronic Pain Grading Scale (39) also asks about functionality using such questions as "In the past 3 months, how much has this pain interfered with your daily activities (e.g. getting dressed, doing shopping)".

Knowledge and Beliefs
Four studies gathered data on patient knowledge and beliefs (15,16,17,25), employing ve different health outcome measures. Briggs et al. (15,33) used the Back Pain Beliefs Questionnaire, which consists of 14 questions exploring beliefs regarding issues such as the management and prognosis of back trouble (45). They also conducted telephone interviews to understand participant's beliefs regarding the aetiology and course of their back pain. MacLeod et al. (40) used the Modi ed Consumer Assessment of Healthcare Providers and Systems survey to assess patient dissatisfaction in areas such as general healthcare and doctors. Farin, Ullrich and Nagl (16) used a single-item measure -"How would you rate your health?" to evaluate participant beliefs on their health status. Finally, Devraj, Herndon and Gri n (17) developed a 12-item survey based on pre-existing pain guidelines and literature to assess the pain awareness and medication knowledge of their participants.

Resource utilisation
Four studies (Glassman et al., Camerini and Schulz, MacLeod et al., Burke, Nahin and Stussman) involved data on resource utilisation (35,(39)(40)(41). A wide variety of resources were studied, and we broadly grouped these outcomes ( Table 4) into utilisation of medications, utilisation of healthcare appointments (e.g. lumbar spine treatment, physiotherapy), and healthcare costs (e.g. expenditure, workdays missed). Of these, utilisation of healthcare appointments was measured the most -in nine studies, while healthcare costs were measured six times, and utilisation of medications was measured thrice (Table 4). Table 5 here. Provide an explicit statement of the question(s) and objective(s) being addressed with reference to their key elements (e.g., population or participants, concepts and context), or other relevant key elements used to conceptualize the review question(s) and/or objective(s)).

Protocol and registration 5
Indicate if a review protocol exists, if and where it can be accessed (e.g., web address), and, if available, provide registration information including registration number.

6-7
Eligibility criteria 6 Specify the characteristics of the sources of evidence (e.g., years considered, language, publication status) used as criteria for eligibility, and provide a rationale.

6
Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with authors to identify additional sources) in the search, as well as the date the most recent search was executed.

5-6
Search 8 Present the full electronic search strategy for at least one database, including any limits used, such that it could be repeated. Describe the methods of charting data from the included sources of evidence (e.g., piloted forms; forms that have been tested by the team before their use, whether data charting was done independently, in duplicate) and any processes for obtaining and con rming data from investigators.

7
Data items 11 List and de ne all variables for which data were sought and any assumptions and simpli cations made. If done, present data on critical appraisal of included sources of evidence (see item 12).

Results of individual sources of evidence 20
For each included source of evidence, present the relevant data that were charted that relate to the review question(s) and objective(s).

21
Summarize and/or present the charting results as they relate to the review question(s) and objective(s).

7-13
Appendix Summarize the main results (including an overview of concepts, themes, and types of evidence available), explain how they relate to the review question(s) and objectives, and consider the relevance to key groups.

13-16
Limitations 25 Discuss the limitations of the scoping review process. 16

Conclusions 26
Provide a general interpretation of the results with respect to the review question(s) and objective(s), as well as potential implications and/or next steps.

16-17
Funding Funding 27 Describe sources of funding for the included sources of evidence, as well as sources of funding for the scoping review. Describe the role of the funders of the scoping review. 3.5 Comparison of included back pain health outcomes against the COS for clinical trials in non-speci c low back pain A total of 50 health outcome measures were utilised across the ten studies reviewed. Of these, 14 (28%) were deemed to be directly related to those in the COS but were limited to two outcomes "pain intensity' and "physical functioning" (18) ( Table 5). The Pain Numerical Rating Scale, Pain Visual Analogue Scale, and Short-form McGill Pain Questionnaire directly addressed the core outcome of "pain intensity", while the ODI, Euro-QOL5D, Chronic Pain Grading Scale, and questions on intrusion of daily and recreational activities (15,33) directly addressed the outcome "physical functioning". Measures on pain duration and frequency were only indirectly related. The COS "health-related quality of life" saw the greatest number of measures indirectly addressing it through outcomes on behaviour, knowledge and beliefs, and resource utilisation since these measures did not directly assess the "impact on physical, psychological and social domains of health" (18). The COS outcome "Number of Deaths" was not explored in any of the included studies.

Association between HL and back pain health outcomes
Although not a primary aim of this scoping review, we brie y detail here ndings on the association between HL and back pain health outcomes ( Table 2): out of six studies analysing the relation between HL and levels of pain and disability, only two found a signi cant association, particularly in the area of pain intensity. On behavioural impact, HL had no signi cant associations with fear avoidance, pain catastrophising, and psychological empowerment. However, patients with low HL were found to have a less active lifestyle. Considering patient knowledge and beliefs, those with lower HL had more di culty identifying types and sources of treatment for back pain and were more dissatis ed with their care. However, no signi cant association was found between HL and beliefs about one's future with back pain. Regarding resource utilisation, it seems that low HL was associated with higher utilisation of curative or symptomatic treatment (e.g. emergency room visits), and lower utilisation of preventive medicine (e.g. u vaccinations).

Summary of overall ndings
The studies included in this scoping review on the association between HL and back pain adopted a wide variety of measures to document HL (nine measures used) and health outcomes (50 measures used).
Although no restrictions were placed on year of publication, all relevant studies were published in 2010 or later, suggesting that interest in the association between HL and back pain is relatively recent. This may re ect HL being a relatively new concept within healthcare (46), and also the growing interest in back pain as it contributes to rapidly rising healthcare expenditure (10,47). For instance, from 1996 through 2013, US expenditure on low back and neck pain rose by an estimated USD 57.2 billion, becoming the third-highest healthcare spending on a single condition in 2013 (47).
Despite no language or country restrictions being placed on the search, most studies were conducted in developed Western countries. Besides demonstrating a dearth of evidence in non-Western countries, this may be a barometer of societal readiness to integrate HL into back pain management. Most pressingly, there is a notable absence of research attempting to draw associations between back pain and HL in developing countries and in Asia, Africa, and South America. This is in keeping with previous epidemiological studies remarking that low back pain monitoring and research is largely restricted to developed nations, while being under-researched in developing countries (20,21), which are more frequently found in Asia, Africa and South America. Alongside this is an increasing recognition of the need to develop and use culturally sensitive HL tools (48).

Comparison with existing literature
Despite expanding the search and using more open search criteria to include seven additional studies, in line with the systematic review by Edward et al. (23), this scoping review highlights the continuing lack of evidence of the relationship between HL and back pain health outcomes.
Few studies incorporated the four outcome domains in the COS (pain intensity, physical functioning, health related quality of life, and number of deaths); only 14 of the 50 health outcome measures used did so. Moreover, these 14 measures were limited to the two core outcome domains of pain intensity and physical functioning (Table 5). This discrepancy suggests divergence of opinions on what is deemed as a key health outcome for people with back pain. This is concerning given that development of the COS incorporated a comprehensive range of views, employing a Delphi process with patients, care providers and researchers, a review by panellists who had published extensively on low back pain, and by a four-continent International Steering Committee (18). Given the rigour and robustness in developing the COS, future studies on back pain are strongly recommended to adopt them to allow for comparison of ndings and development of a rigorous evidence-base. Furthermore, the authors of the COS have subsequently argued for the inclusion of the 24-item Roland-Morris Disability Questionnaire for physical functioning, and the Short Form Health Survey 12 and 10-item PROMIS Global Health form for health-related quality of life (49). However, none of these tools were used in the included studies (Table 2).

Implications for future research and practice
Several limitations were noted in the literature with implications for future research design, speci cally regarding study design, measures used and included study populations: All ten studies were cross-sectional with no studies exploring the longitudinal relationship between HL and back pain outcomes, nor the e cacy and implementation of HL interventions for people with back pain. Although a mixed-methods approach is preferable to holistically evaluate the complex construct of HL (50), only the 2010 study by Briggs et al. employed quantitative and qualitative approaches (15).
Another limitation of in terms of design was that the primary studies relied heavily on patient reported outcome measures (PROMs), which may be biased by one's physical and psychological states, along with one's memory, willingness and ability to answer the questions. This may in uence one's ability to give accurate self-assessments of health status (51). This limitation could be overcome by the concurrent use of objective markers (e.g. functional tests), diagnostic imaging (e.g. functional magnetic resonance imaging), and/or observer reported outcomes (52,53).
Many studies also had limitations in terms of the HL measure used. Most HL measures employed assessed functional HL, while only a minority assessed communicative and critical HL. For example, both the S-TOFHLA and NVS focuses solely on functional HL, speci cally numeracy and reading comprehension skills, which are merely sub-components of "understanding health information" in the de nition of HL. The HeLMS provides a more holistic measure of HL, for instance by asking questions on the assessment and usage of health information. HeLMS, and the questionnaires used by Camerini and Schulz (39), MacLeod et al. (40) and Burke, Nahin and Stussman (41) have not been used as widely as the S-TOFHLA and NVS, and their content validity in other settings requires con rmation. While a varied questioning style is likely to result in a more complete assessment of HL, measures tended to focus purely on either objective response (e.g. S-TOHHLA and NVS) or subjective replies (e.g. HeLMS). Ideally future studies on HL should use measures that cover the three domains of HL as well as have vigorous validation in the setting employed.
Study population characteristics were also a source of limitation. Briggs et al. (15,33) faced a limited distribution of HL, hampering efforts to analyse relationships by HL level. Five of the included studies excluded patients based on their language literacy, effectively excluding lower HL participants. If basic language pro ciency is required to obtain selfreported patient outcomes, this comes at the cost of excluding certain sectors of the population. The use of translators or pictorial questionnaires needs to explored to enable the inclusion of participants who may be considered vulnerable, for example those facing communication barriers or multimorbidity (54,55).
Responder bias through self-selection was another common limitation in terms of study population design. This is important in the context of HL studies, as low HL patients with lesser ability to communicate well with their healthcare provider may have a tendency to decline study involvement (56). This limitation may potentially be mitigated using retrospective and anonymised data, rather than depend on the voluntary actions of patients.

Strengths and limitations of this review
The strengths of this review include the wide search, involving 13 bibliographic databases with no search limiters or lters. By including studies on all forms of back pain health outcomes associated with HL, this review was able to build on the preliminary work of Edward et al.
The studies by Devraj, Herndon and Gri n (17), MacLeod et al. (40), Farin, Ullrich and Nagl (16), and Köppen et al. (8) did not have a solely back pain population, but were included as the back pain population made up at least 10% of the overall study, which was pre-determined as the cut-off percentage for eligibility into this review. This cut off has been used as a marker for signi cance in other studies (57,58), but caution may be needed when interpreting the results of these studies.
Due to the heterogeneity of HL measures and back pain outcome measures employed, as well as the paucity of relevant studies, it was di cult to compare the ndings across studies and provide rm conclusions on the association between HL and each back pain health outcome.

Conclusions
Overall, the ten relevant studies included in this review yielded a total of nine different measures of HL and 50 measures of back pain health outcomes. Also, most health outcomes evaluated by the included studies did not seek to directly satisfy the core outcome domains for clinical trials in non-speci c low back pain. The wide variety of measures used hampers efforts to form conclusive relationships between HL and the respective health outcomes and precludes the use of a meta-analysis approach. Furthermore, while exploring a wide range of health outcomes, studies overlooked the core outcome domains for clinical trials in non-speci c low back pain.
Even with an extensive search of the literature only ten relevant studies were identi ed. In Asian developing nations, where back pain related morbidity is high, no relevant studies were found, highlighting the need for more culturally speci c research into HL and its relationship with back pain health outcomes. Observing cultural nuances while respecting the need to streamline the study of HL by standardising the outcome measures used is challenging, but it is not impossible as evidenced by the HLS-EU-Q47, which has been validated in six Asian countries since being developed for European use (59).
In summary, to allow for comparison across ndings and the development of a rigorous evidence base, future works should seek to address the core outcome set for clinical trials in non-speci c low back pain. Furthermore, research thus far has focused on a narrow range of populations and there is an urgent need to broaden the evidence-base to include those with characteristics known to be moderators in health outcomes. Such work demands the incorporation of comprehensive measures of health literacy that have both generic and culturally sensitive components. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
Not applicable.
Authors contributions and acknowledgements are included in the title page for the purposes of blind review.
Authors' contributions CS provided review on titles and abstracts, full-text reviews of the shortlisted papers, and was a major contributor in writing the manuscript; BB and HES were involved in project conception, analysis and manuscript writing; LTC contributed to analysis and manuscript writing. JP contributed to manuscript writing and acted as the eld expert in the consultation process. WWC collected the raw data, provided review on titles and abstracts, and full-text reviews of the shortlisted papers, and contributed to manuscript writing. All authors read and approved the nal manuscript. Preferred Reporting Items for Systematic Reviews and Meta-Analyses ow diagram of the literature review.