Patient-centred outcomes in lateral elbow tendinopathy: A systematic review of available evidence in UK populations

Abstract

Background

For Patient-Reported Outcome Measures to be useful, they must have evidenced reliability, validity and responsiveness in both the condition and population of interest. The aim of this study is to systematically review the evidence for Patient-Reported Outcome Measures in UK patients with lateral elbow tendinopathy.

Methods

A systematic search was performed in Ovid MEDLINE, Embase and CINAHL. Studies were included if reporting administration of Patient-Reported Outcome Measures in UK populations with lateral elbow tendinopathy. Patient-Reported Outcome Measures characteristics and target populations were assessed using a structured classification system. Patient-Reported Outcome Measures reporting in randomised controlled trials was assessed against Consolidated Standards of Reporting Trials (CONSORT) standards (Patient-Reported Outcome extension).

Results

A total of 16 articles were included. Out of seven different Patient-Reported Outcome Measures, there was evidence of partial validation for five. The assessment of validity, reliability and responsiveness of all Patient-Reported Outcome Measures in lateral elbow tendinopathy UK populations extended to just 20 individual patients. No articles conformed to the CONSORT Patient-Reported Outcome extension standards.

Conclusion

There exists a paucity of data on the psychometrics of Patient-Reported Outcome Measures in UK lateral elbow tendinopathy populations. Without these data, trial design and interpretation are significantly hindered. The high prevalence of this condition and significant volume of studies being conducted into novel treatments highlight the need for this knowledge gap to be resolved.

Background

Lateral elbow tendinopathy (LET), known more commonly as tennis elbow, is a prevalent and potentially debilitating condition.^1,2 Though the condition is regarded as benign and self-limiting, absenteeism due to LET in the UK is estimated to cost the economy £27 million per annum.³ With a UK prevalence between 1.5 and 3%,² it is surprising that no clear treatment consensus exists.^4,5 This clinical variability has driven a large volume of research activity, with over 80 registered trials currently ongoing.⁶ However, to be confident in our treatments, we must be certain that the outcome measures used in these trials truly reflect patient benefit or harm.

Successful treatment in LET can be regarded as amelioration of pain and return of function. Constructs such as these are now commonly quantified through the use of Patient-Reported Outcome Measures (PROMs). Collaborative work by academics and clinicians has crystallised in the development of systematic, robust and valid ways of collecting health outcomes from patients that purport to quantify, in a meaningful way, how the patient feels their condition affects them.⁷ In reference to musculoskeletal pathology, this has resulted in numerous PROMs used to quantify the burden of a specific disease, such as the use of the Oxford Hip and Knee scores in assessing the outcome of joint arthroplasty.⁸

Appropriate outcome measures must demonstrate that they are acceptable to patients, reliable, valid and responsive (sensitive to change).⁹ When the outcome measure has been developed in a different clinical or geographical population, there needs to be evidence of equivalence both in a disease-specific and cross-cultural context.^10,11 The validation of a PROMs in a new language is accepted as integral to its interpretability, what may be less appreciated is the requirement for cultural adaptation, for a PROM delivered in its original language but in a different geographical location and in a different population.¹² Though not previously investigated in elbow pathology, it is notable that a recent evaluation of population norms of English language shoulder outcome measures, the Association of Shoulder and Elbow Surgeons shoulder score (University of California, Los Angeles), UCLA shoulder rating scale and Shoulder Pain and Disability Index, between Canada and Australia reported statistically significant variability in the scores between the two countries.¹³

A structured assessment of outcome measurement in LET in UK populations has not been undertaken. This study aimed to address this gap by systematically assessing the outcome measures used for measuring PROMs in LET in a UK population and to assess the reporting of randomised controlled trials (RCTs) using PROMs in LET. Only when valid outcomes have been identified, can recommendations on choice of outcome measures for future research be made.

Materials and methods

PRISMA guidelines on the reporting of systematic reviews were followed.¹⁴ All articles reporting the development, psychometric evaluation or use of PROMs in LET in UK adults (≥18 years) were included. Any measures of symptoms and functioning in LETs that involved a patient-reported outcome measurement (regardless of whether this also contained a physician-reported outcome component) were included. Studies in paediatric populations, case reports, case studies and conference abstracts were excluded.

A search strategy was constructed using MeSH and free-text terms (online Appendix 1). The search strategy development was guided by previously published search strategies for systematic reviews of interventions in elbow pathology¹⁵ and for the identification of outcome measures,¹⁶ along with terms specifically selected in order to capture names of relevant instruments published in previous systematic reviews of elbow-specific rating scales.^17–20 The strategy was further adapted to each database through the modification of thesaurus terms, wildcards and truncation. The search was run on 1 May 2017 in Medline (Ovid MEDLINE, 1948–2016 and Ovid MEDLINE In-Process & Non-indexed Citations) accessed through OVIDSP, Embase (Embase 1974–2016) accessed through OVIDSP and CINHAL (CINHAL 1981–2016) accessed through EBSCO host.

The screening process was conducted in a stepwise manner. At each stage, one researcher (JPE) and a further researcher reviewed each title and abstract. In cases of disagreement, the article proceeded to the next stage of review to ensure maximum sensitivity.

At the full-text level, articles were also subcategorised in two groups: articles reporting primary research on the development and/or psychometric evaluation of PROMs in LET in UK populations (development) and articles reporting the use of outcome measures in clinical studies in UK populations (use).

Data synthesis

Development articles were classified according to three guiding concepts, using the structured classification system proposed by Valderas and Alonso²¹: construct (the measurement object), population (based on age, gender, condition and culture) and measurement model (dimensionality, metric and adaptability).²¹

The assessment of construct denotes, for the purpose of this study, the range of characteristics measured by the outcome measure, which are affected by LET. The construct analysis has, at its foundation, the conceptual strengths of the Wilson and Cleary model,²² but is also integrated with the theoretical model that underpins the International Classification of Functioning, Disability and Health. A strength of the model that is particularly pertinent in the assessment of LET outcome measures is the systematic consideration of intended population of use. Within the axis of population consideration of culture is also made, where there is information pertaining to the dyad of language and country for which the outcome measures have been devised.

It should be noted that this system is only descriptive and does not provide any fundamental evaluation of measurement properties.²¹ But in the early stages of outcome measure assessment, this approach provides the clearest method of identifying the candidate pool of measures. Only once this is undertaken and deemed to be adequate, can a systematic evaluation of measurement properties in a specific population of use be undertaken.

Articles reporting the use of PROMs (use) were peer-reviewed, published articles with outcome measure evaluation in a population of LET patients. Date of publication, outcome measure(s) chosen and population of use were extracted. For randomised control trials, the CONSORT Patient-Reported Outcome (PRO) extension²³ was used to systematically assess the reporting of outcome measure choice and justification. The original CONSORT statement aims to encourage transparent and complete reporting of clinical trials and is associated with improved reporting practice.²⁴

An a priori hypothesis was formulated with regard to informed choice of outcome measures in UK populations. We hypothesised that articles reporting the use of PROMs would more frequently use PROMs for which there would be evidence from studies of validation of such measures in UK populations.

Results

We identified 7261 records from the electronic database search. A total of 16 articles met the inclusion criteria: five articles reporting the development and/or psychometric evaluation of outcome measures in LET-specific patients and 11 articles reporting their use in a UK population (Figure 1) (online Appendix 2).

Figure 1.

PRISMA flow chart of the systematic literature review. LET: lateral elbow tendinopathy; PROMs: Patient-Reported Outcome Measures.

Measures

Five outcome measures were identified that were developed or had undergone psychometric evaluation, on UK populations that at least, in part, contain patients with LET (Table 1). They were all fully standardised measures that had all been developed for measuring symptoms (mainly pain) and functioning in English-speaking UK adults of either gender. However, only one of them, the Patient-Rated Tennis Elbow Evaluation (PRTEE) was LET specific, the remaining instruments were developed as elbow-specific tools designed for varying pathologies but including in their validation a subsample of LET patients. Two outcome measures were originally developed for UK populations: the Oxford Elbow Score (OES) and the Liverpool Elbow Score (LES). The remaining three outcome measures were developed in the English language outside of the UK (US, Canada and Australia) but had undergone some level of psychometric evaluation in UK populations. Of note, no modification was deemed necessary in the wording or description of the symptoms or activities measured for any of those instruments.

Table 1.

Outcome measures for the assessment in lateral elbow tendinopathy (LET) with psychometric evaluation in UK population.

Outcome measure	Country of origin	Exclusively patient reported (no. items)	Construct (no. items)	Populationa	Measurement modelb	UK LET assessment
Oxford Elbow Score (OES)23	United Kingdom	Yes (12)	A1. Symptoms Pain (4) A2. Function Elbow function (4) Psychosocial (4)	Assessment of outcome of surgery of the elbow23	C1. Profile C2. Psychometric C3. Completely Standardised	Surgically treated LET patients make up 11.2% (n = 12/107) of the total development and validation cohort23,32,33
Liverpool Elbow Score (LES)24	United Kingdom	No, physician administered (15)	A1. Symptoms Pain (1) A2. Function Range of motion (4) Strength (1) Ulnar nerve function (1) Activity (8)	B1. Adults B2. All genders B3. Assessment of elbow pathology in tertiary care setting24 B4. UK English	C1. Index C2. Psychometric C3. Completely Standardised	Tertiary care patients with LET make up 12.7% (n = 8/63) of the total development and validation cohort24
Patient-rated Tennis Elbow Evaluation (PRTEE)34	Canada	Yes (15)	A1. Symptoms Pain (5) A2. Function Activity (10)	B1. Adults B2. All genders B3. Lateral Elbow Tendinopathy patients34 B4. UK English	C1. Index C2. Psychometric C3. Completely Standardised	57 LET patients (100% of cohort)25 (PRTEE delivered in a modified form but not formally cross-culturally validated)
Disabilities of the Arm, Shoulder and Hand (DASH)35 2 × Optionalmodules Work sporting/performing arts	US, Canada, Australia	Yes (30)	A1. Symptoms Pain (5) A2. Function Physical function (21) Psychosocial (4)	B1. Adults B2. All genders B3. Applied to multiple elbow pathologies36 B4. UK English	C1. Index C2. Psychometric C3. Completely Standardised	Surgically treated LET patients make up 11.2% (n = 12/107) of the total development and validation cohort23,32,33 Tertiary care patients with LET make up 12.7% (n = 8/63) of the total development and validation cohort24 (DASH delivered in original form, without any modifications) (UK English DASH translation available from 201528)
Mayo Elbow Performance Score (MEPS)37 Physician administered eight items: 1× pain 1× Range of motion 1× Instability 5× Function	United States	No, physician administered (15)o	A1. Symptoms A2. Function	B1. Adults B2. All genders B3. Applied to multiple elbow pathologies38 B4. UK English	C1. Index C2. Clinometric C3. Completely standardised	Surgically treated LET patients make up 11.2% (n = 12/107) of the total development and validation cohort23,32,33 (MEPS delivered in original form, without any modifications)

^aAll measures were developed for English-speaking adults of either gender.

^bAll measures were fully standardised.

Only the PRTEE has had its metric properties assessed in a UK cohort that was exclusively diagnosed with LET. This was conducted on 57 patients to quantify the Minimal Clinically Important Difference (MCID) of the PRTEE. This study formed part of a larger prospective trial assessing microcurrent therapy in LET and analysed data from 57 individuals with clinically and sonographically diagnosed LET who all underwent microcurrent therapy. They report a weak correlation between the PRTEE and global change scale, but no assessment of construct validity or any other metric assessment is undertaken. For the four remaining outcome measures, the proportion of patients included within their study cohorts who were diagnosed of LET ranged from 11 to 12.7% (Table 1). None were evaluated in more than 12 patients, and as multiple measures were reported on the same patient’s cohorts, when all individual patients from these studies were tallied it reveals that this equates to 20 UK LET patients in total.

Eleven additional articles reported using PROMs to evaluate disease impact in UK populations with LET. These studies were published between 2003 and 2014 (Table 2). Out of the five outcome measures for which there had been a previous psychometric evaluation, only three were subsequently applied to evaluate LET outcomes (DASH, OES and PRTEE). The outcome measures that were not utilised were the LES and the Mayo Elbow Performance Score (MEPS). Perhaps more surprisingly, two additional measures were used, namely the Nirschl score and the Patient-Rated Wrist Evaluation (PRWE), although no evidence on the psychometric properties or even their cross-cultural equivalence was available. Overall, the PRTEE (and precursor PRFEQ) was reported six times, the DASH four times, the Nirschl score twice, the OES once and the PRWE once. Seven of the 11 studies stated that the outcome measure was their study’s primary outcome, where the need for clear and evidenced validity is at its greatest.

Table 2.

Studies reporting the use of PROMs in patients with lateral elbow tendinopathy.

Author	Year	Title	Study type and population	Outcome measure
Melikyan et al.28	2003	Extracorporeal shockwave treatment for tennis elbow: a randomised double-blind study	RCT LET patients who failed conservative treatment	DASH
Dunkow et al.27	2004	A comparison of open and percutaneous techniques in the surgical treatment of tennis elbow	RCT LET patients who failed conservative treatment	DASHa
Connell et al.37	2006	Ultrasound-guided autologous blood injection for tennis elbow	Prospective cohort LET patients who failed conservative treatment	Nirschla
Alizadehkhaiyat et al.38	2007	Pain, functional disability, and psychologic status in tennis elbow	Cross-sectional LET with symptoms lasting > 3 months	DASH PRWE PRFEQ
Connell et al.39	2009	Treatment of lateral epicondylitis using skin-derived tenocyte-like cells	Prospective pilot study (not randomised) LET patients who failed conservative treatment	PRTEEa
Clarke et al.40	2010	Lateral elbow tendinopathy: correlation of ultrasound findings with pain and functional disability	Prospective cohort LET who had not undergone invasive treatment	PRTEEa
Creaney et al.25	2011	Growth factor-based therapies provide additional benefit beyond physical therapy in resistant elbow tendinopathy: a prospective, single-blind, randomised trial of autologous blood injections versus platelet-rich plasma injections	RCT LET patients who failed conservative treatment	PRTEEa
Nazar et al.41	2012	Percutaneous tennis elbow release under local anaesthesia	Prospective cohort LET patients who failed conservative treatment	DASHa OES
Stenhouse et al.42	2013	Do blood growth factors offer additional benefit in refractory lateral epicondylitis? A prospective randomised pilot trial of dry needling as a stand-alone procedure versus dry needling and autologous conditioned plasma	Prospective pilot study (randomised) LET patients who failed conservative treatment	Nirschl
Maffulli et al.43	2014	Assessment of the Effectiveness of Extracorporeal Shock Wave Therapy (ESWT) For Soft Tissue Injuries (ASSERT): An online database protocol	Online database protocol of clinically or radiologically confirmed LET	PRTEEa
Tonks et al.26	2007	Steroid injection therapy is the best conservative treatment for lateral epicondylitis: a prospective randomised controlled trial	RCT LET patients who had not had treatment for the preceding six months	PRTEE

DASH: Disabilities of the Arm, Shoulder and Hand; LET: lateral elbow tendinopathy; OES: Oxford Elbow Score; PRFEQ: Patient-Rated Forearm Evaluation Questionnaire; PROM: Patient-Reported Outcome Measure; PRTEE: Patient-rated Tennis Elbow Evaluation; PRWE: Patient-Rated Wrist Evaluation; RCT: randomised controlled trial.

^aPrimary outcome.

Four of these 11 studies were RCTs.^25–28 The level of adherence to CONSORT standards for reporting PROMs outcomes for RCTS for the four trials suggested substantial room for improvement (Table 3). No information was available for three standards for any RCT and only partial information was available for the other two standards.

Table 3.

Adherence to CONSORT reporting standards (PRO extension) of UK-based lateral elbow tendinopathy randomised controlled trials.

CONSORT 2010 statement	PRO extension	Studies meeting the requirements
Structured summary of trial design, methods, results and conclusions	The PRO should be identified in the abstract as a primary or secondary outcome	1/425
Specific objectives or hypotheses	The PRO hypothesis should be stated and relevant domains identified, if applicable	0/4
Completely defined pre-specified primary and secondary outcome measures, including how and when they were assessed	Evidence of PRO instrument validity and reliability should be provided or cited if available, including the person completing the PRO and methods of data collection (paper, telephone, electronic, other)	0/4 (validity of PROM in UK population) versus 2/4 (validity of PROM in another LET population)25,261/4 (data collection method)25
Statistical methods used to compare groups for primary and secondary outcomes	Statistical approaches for dealing with missing data are explicitly stated	0/4
Trial limitations addressing sources of potential bias, imprecision, and, if relevant multiplicity of analyses	PRO-specific limitations and implications for generalisability and clinical practice should be discussed	0/4

CONSORT: Consolidated Standards of Reporting Trials; LET: lateral elbow tendinopathy; PRO: Patient-Reported Outcome; PROM: Patient-Reported Outcome Measure.

Discussion

This study has identified a lack of evidence with which to inform outcome measure choice in LET in the UK. Future validation of outcome measures in UK populations is required in order to be able to ground any recommendations on a firm evidence base. Furthermore, some outcome measures are currently being used as primary outcomes in UK-based studies in the absence of any evidence for their cross-cultural appropriateness and psychometric properties.

We were able to retrieve at least some evidence of the evaluation of the psychometric properties of five outcome measures. The PRTEE is the only measure specifically designed for the evaluation of a LET population. All measures attempt to measure the domains of function and symptoms in adults. All but the DASH have been designed to assess these domains in reference to the elbow exclusively.

The total reporting of validity, reliability or reproducibility of outcome measures in UK LET patients is limited to 20 patients.^29,30 All of these patients have been embedded in larger cohorts containing a heterogeneous group of elbow pathology. Due to the limited size of this LET sample, it has been unfeasible for the reporting authors to conduct a standardised psychometric assessment of the outcome measures using methods such as COSMIN or EMPRO.

The largest assessment outcome measure utility in UK LET patients was published by Poltawski and Watson³¹ and included 57 patients. Although this is by far the largest sample of LET patients of any of the studies included here, outcome interpretability through derivation of MCID score was undertaken with no evaluation of other relevant psychometric characteristics. The PRTEE was not originally designed for a UK population and no evidence of formalised cross-cultural evaluation is presented. This would always be necessary when applying a new instrument to a different population, as the use of language across continents, though English in origin, confers both linguistic and cultural differences. In upper limb pathology this has been recently demonstrated but in this case the need was additionally increased by the fact that items in the PRTEE had been altered prior to administration (the words coffee and milk were removed from the item ‘Lift a full coffee cup or glass of milk to your mouth’, ‘pants’ were replaced by ‘trousers’ and ‘washcloth or wet towel’ by ‘wet cloth’). The authors acknowledge that the altering of the outcome measure wording may have altered its measurement properties.³¹

In many circumstances it will be completely appropriate and even highly advisable to alter the wording of outcome measures. However, it should be undertaken under the principles of cross-cultural adaptation.^10,12 It is widely recognised that if a measure is to be used across cultures, the items must be both linguistically translated and culturally adapted to maintain the content validity of the outcome measure at a conceptual level.¹² Guillemin et al.³² have proposed scenarios that should alert authors to situations where translation or adaptation should be undertaken. In the situation of an outcome measure being used in another country, but in the same language, cultural adaptation is required. For LET in UK populations, this would be the case for the DASH, MEPS, PRTEE and Nirschl outcome measures. Of note, the DASH and quick DASH score have been culturally adapted to UK English since 2015.³³ To the best of our knowledge, this score had not been utilised in any of the identified studies.

The process of cross-cultural adaptation has been well reported.¹⁰ A 10-stage process proposed by the International Society for Pharmacoeconomics and Outcomes Research¹⁰ involves forwards and backwards adaptation by multiple reviewers, cognitive interviewing with patient populations and pre-testing of the final questionnaire. Though this may be seen as a laborious process, users of measures that have not been rigorously adapted must also be aware that language alterations may alter measurement properties. Therefore, reference values for group comparison, minimally important difference data or power calculations may not be valid in the new cultural context of use.

This study has identified that the reporting of outcome measures in UK LET RCTs does not conform to the CONSORT-PRO guidance. Though two of the studies were published prior to the guidance publication in 2010, the stark paucity of reporting of outcome measure detail is concerning. This lack of reporting is in line with the deficits in outcome measure validity highlighted through the Valderas and Alonso²¹ classification system. Though we hypothesised that there would be a preference for outcome measures with published validity in the target population, we have identified that with the current level of evidence this is not possible. This lack of suitable outcome measures has been identified by other authors.^34,35 Long et al.⁴ reported in their National Institute of Health Research, Health Technology Assessment review of systematic reviews of conservative treatments in LET that a lack of standardised outcome measures hindered interpretation and synthesis of results.They recommend that the inclusion of a patient-reported measure of upper extremity function in interventional trials would ease results synthesis. However, we have identified that the lack of a clear choice within the UK population is likely to significantly hinder a researcher’s ability to undertake this.

The authors acknowledge the inherent limitations of this study. The search strategy may have failed to identify all outcome measures used, and the identification of the study populations’ nationality in interventional trials can be prone to error. However, attempts were made to ensure that the strategy was as robust as possible.Outcomes in LET can be measured in numerous ways, including grip strength, pain provocation tests and visual analogue scales to mention a few, and although this may be a highly legitimate method and was not assessed as part of this study. The authors feel that this approach is justified owing to the increasing view that the ultimate measure of success in health care is whether it helps patients from their own point of view.³⁶ Outcome measures, that quantify patient’s health-related quality of life, with particular reference to PROMs, are recommended by National bodies across the world, including the NIHR in the UK and FDA in the USA.⁷ Furthermore, the use of condition-specific PROMs is increasingly common in musculoskeletal medicine and are collected as part of the English NHS PROMs programme.⁷ With the increasing use of PROMs used as primary outcomes in clinical trials, it is, therefore, relevant that their use is rigorously assessed.

This study has identified that, with current levels of evidence, it is not appropriate to recommend any PROMs for LET studies in UK populations. Though the OES, PRTEE and DASH show potential as patient-reported measures, with domains likely to be appropriate in LET, further assessment is required in UK populations to quantify their validity, reliability, responsiveness and patient acceptability.

Take-home messages

There is some evidence for the psychometric properties of OES, PRTEE and DASH PROMs in the assessment of patients with LET. Robust evidence on the validity, reliability and responsiveness of any PROM in UK populations of LET patients is lacking.

Authors’ note

The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care South West Peninsula (NIHR CLAHRC South West Peninsula).

Ethical Review and Patient Consent

Not required for this article.

Supplementary material

Appendix 1 and Appendix 2 are available as supplementary material at: journals.sagepub.com/doi/suppl/10.1177/1758573218791164.