Article Text

Download PDFPDF

Psychometric properties of knee proprioception tests targeting healthy individuals and those with anterior cruciate ligament injury managed with or without reconstruction: a systematic review protocol
  1. Ashokan Arumugam1,
  2. Andrew Strong1,
  3. Eva Tengman1,
  4. Ulrik Röijezon2,
  5. Charlotte K Häger1
  1. 1 Department of Community Medicine and Rehabilitation – Physiotherapy Section, Umeå University, Umeå, Sweden
  2. 2 Department of Health Sciences – Physiotherapy Section, Luleå University of Technology, Luleå, Sweden
  1. Correspondence to Dr Ashokan Arumugam; ashokanpt{at}


Introduction An anterior cruciate ligament (ACL) injury affects knee proprioception and sensorimotor control and might contribute to an increased risk of a second ACL injury and secondary knee osteoarthritis. Therefore, there is a growing need for valid, reliable and responsive knee proprioception tests. No previous study has comprehensively reviewed all the relevant psychometric properties (PMPs) of these tests together. The aim of this review protocol is to narrate the steps involved in synthesising the evidence for the PMPs of specific knee proprioception tests among individuals with an ACL injury and knee-healthy controls.

Methods and analysis The Preferred Reporting Items for Systematic reviews and Meta-Analyses will be followed to report the review. A combination of four conceptual groups of terms—(1) construct (knee proprioception), (2) target population (healthy individuals and those with an ACL injury managed conservatively or with a surgical reconstruction), (3) measurement instrument (specific knee proprioception tests) and (4) PMPs (reliability, validity and responsiveness)—will be used for electronic databases search. PubMed, AMED, CINAHL, SPORTDiscus, Web of Science, Scopus, the Cochrane Central Register of Controlled Trials and ProQuest will be searched from their inception to November 2018. Two reviewers will independently screen titles, abstracts and full text articles, extract data and perform risk of bias assessment using the updated COnsensus-based Standards for the selection of health Measurement INstruments risk of bias checklist for the eligible studies. A narrative synthesis of the findings and a meta-analysis will be attempted as appropriate. Each PMP of knee proprioception tests will be classified as ‘sufficient’, ‘indeterminate’ or ‘insufficient’. The overall level of evidence will be ascertained using an established set of criteria.

Ethics and dissemination Ethical approval or patient consent is not required for a systematic review. The review findings will be submitted as a series of manuscripts for peer-review and publication in scientific journals.

PROSPERO registration number CRD42018108014.

  • knee
  • musculoskeletal disorders
  • orthopaedic sports trauma
  • rehabilitation medicine
  • sports medicine

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

View Full Text

Statistics from

Strengths and limitations of the study

  • A comprehensive systematic review of the psychometric properties of specific knee proprioception tests, following the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines, using a broad search in several electronic databases.

  • The corroboration of evidence using an established set of criteria that relies on the scores of the updated COnsensus-based Standards for the selection of health Measurement INstruments risk of bias evaluation of the eligible studies and the quality of psychometric properties graded with a standard set of criteria.

  • The review provides an up-to-date compilation of current knowledge about the psychometric properties of specific objective knee proprioception tests, and identification of eventual need for further research.

  • The review is limited to original research articles published in English.

  • The review is limited to studies investigating adolescents (≥10 years) and adults with an ACL injury (with or without reconstruction) and/or healthy participants.


Anterior cruciate ligament (ACL) ruptures are one of the most common knee injuries among athletes.1 Annual incidence rates among amateur athletes range from 0.03% to 1.62%, with these figures rising further among professional athletes from 0.15% to 3.67%.2 Treatment involves physiotherapy with or without surgery. Despite treatment, return-to-sport at pre-injury level is only achieved by approximately 65% of individuals with an ACL injury.3 Those who do return-to-sport are at a six times higher risk for an ACL injury than non-injured individuals.4 Thus, despite completing rehabilitative interventions, the short-term consequences of an ACL injury include a reduced level of physical activity and increased risk for further ACL injury. The long-term consequences following an ACL injury also include the more than 50% chance of developing knee osteoarthritis.5 An ACL injury therefore negatively affects the short-term and long-term health of affected individuals, which in turn places a burden on healthcare systems.

Proprioceptors such as Ruffini nerve endings, Pacini receptors and Golgi tendon organ-like endings are present in the intact/injured ACL.6 7 These receptors provide afferent feedback to the central nervous system regarding the sensation of limb position and movement as well as the senses of force and effort.8 Injury to the ACL causes damage to and loss of these proprioceptors9 10 and, for instance, affects muscle spindle excitability of the thigh muscles.11–13 Reduced proprioception is evident in ACL-injured knees as compared with the contralateral intact knee of the same individuals as well as to knee-healthy controls.14 15 Such reduced proprioception/sensorimotor control is believed to be a major contributing factor to the 30–40 times increased risk of a second ACL injury16 and post-ACL injury knee osteoarthritis development.17 Identifying both knee-healthy individuals and those having an ACL injury with poor knee proprioception may thus aid the prescription of targeted neuromuscular training, which could facilitate injury/re-injury prevention strategies as well as short-term and long-term rehabilitation outcomes.

Assessing knee proprioception is currently performed in a number of ways, with each test aimed at isolating one of the proprioceptive senses. The most common methods strive to determine joint position sense (JPS) and threshold to detect passive motion (TTDPM).18 JPS typically requires the attempted matching of a target knee joint angle either ipsilaterally or contralaterally using an active and/or a passive movement.19 20 Test outcome is the difference in degrees between the target angle and the attempted reproduction angle. TTDPM most often involves blindfolded individuals signalling the onset of passive movement from a pre-set joint angle and also identifying the movement direction.21 Test outcome is the time between actual movement onset and the signal provided as well as whether the correct movement direction is detected by the individual. Other tests targeting proprioception such as force sense (the ability to accurately reproduce forces and/or hold a force steadily for a brief period [e.g. 5 s]),19 21 22 force perception/load identification (the ability to differentiate between different loads/weights),23 velocity sense (the ability to actively reproduce the velocity of a passive movement),21 active movement extent discrimination (AMEDA, the ability to discriminate between two or more active movements of different ranges of motion)24 or psychophysical threshold methods (the ability to detect and discriminate between different joint positions following passive movements)25 are also reported. However, it appears important to differentiate tests based on the specific targeted sense of proprioception due to the reported lack of correlation between knee position sense, motion sense and force sense.26 Regardless of the targeted proprioceptive sense, there are many factors to be considered when designing a knee joint proprioception test such as determining body position, whether the knee should be weight-bearing or non-weight-bearing, knee angles and speed of movement, occluding visual input, restricting other somatosensory information and minimising extraneous variables. Further, the lack of data regarding the psychometric properties (PMPs) of existing knee proprioception tests challenge their utility and clinical application.15 To date, no gold standard knee proprioception tests exist to guide clinicians or researchers to advocate their use.

Previous reviews of the literature have claimed that proprioception tests show uncertainty for their PMPs. Hillier et al. (2015)27 failed to find a proprioception test with well substantiated PMPs and, therefore, questioned their clinical application. On the other hand, Han and colleagues (2016)28 concluded that although JPS tests may be efficient, they have low testing validity owing to differences in the proprioceptive information perceived during (passive) target angle generation and (active) reproduction. Moreover, Smith and colleagues (2013)29 found variable reliability of JPS tests and highlighted the limited evidence for the PMPs of such tests. Both JPS and TTDPM may have less ecological validity because they do not represent function during normal tasks.28 30 On the other hand, AMEDA (ability to discriminate between two or more movements) has been claimed by Han et al. (2016)28 to have (relatively) better ecological and testing validity. Despite the uncertain validity of these tests, they have been widely applied. The criterion and construct validity, reliability and responsiveness of these tests should be systematically evaluated. Moreover, previous reviews either systematically appraised only a few PMPs of one or more knee proprioception tests but not all of them29 or did not corroborate the relevant literature systematically.28 30 To date, neither a systematic review nor a meta-analysis has assimilated the data, as a whole, on all the relevant PMPs of established specific knee proprioception tests in healthy individuals nor those with ACL injuries.

Thus, a systematic review of the literature, implementing updated methodological quality assessment tools such as the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) risk of bias checklist,31 covering the PMPs of all documented proprioception tests specifically targeting the knee is lacking and is warranted. The current protocol narrates the study methods and reporting process of such an ongoing systematic review (and meta-analysis) aimed at corroborating the levels of evidence underpinning the PMPs of existing specific knee proprioception tests among individuals with an ACL injury managed conservatively or surgically and knee-healthy controls.



The systematic review will follow the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines for conducting and reporting the review.32 33

Eligibility criteria

Inclusion criteria

Studies will be included if they meet the following criteria:

  1. Participants: Adolescents (≥10 years of age) and adults who are healthy as well as those with anterior cruciate ligament injury managed with or without reconstruction.

  2. Construct: One or more specific methods measuring joint proprioception such as active or passive JPS (ipsilateral or contralateral matching,19 20 verbal identification of joint position,23 and identifying location of joint motion using contralateral limb),23 TTDPM (low21 or high velocity30), force sense,19 21 force perception,23 velocity sense,21 AMEDA24 or psychophysical threshold methods25 designed for the knee.

  3. Equipment: Any equipment that is readily available or customised for the purpose of quantifying knee joint proprioception (e.g. video cameras, two-dimensional/three-dimensional motion analysis system, electro-goniometer, isokinetic dynamometer, etc.).

  4. Setting: The test procedure executed in a laboratory, clinical or any setting.

  5. Outcome measures: Studies investigating any of the following PMPs of knee proprioception tests—reliability (absolute reliability (agreement) and relative reliability of test–retest, intra-rater and/or inter-rater designs), validity (criterion (concurrent or predictive), construct (hypothesis testing—known-groups or comparison with other outcome measurements), and responsiveness (table 1).

  6. Study type: (1) Studies investigating PMPs of knee joint proprioception tests as one of the (primary) aims or the sole aim of their study, (2) if studies have reported reliability, validity and/or responsiveness as secondary or additional findings, then the full text of these studies will be reviewed and included only if adequate details to rate their quality/risk of bias are available, (3) if studies have included knee proprioception data of individuals with ACL injury and other lower limb disorders and knee-healthy controls, then these studies will be included only if data were reported separately for each group and (4) peer-reviewed observational studies, cross-sectional studies, randomised controlled trials, controlled clinical trials or quasi-experimental studies will be included if one or more PMPs of a specific knee proprioception test have been addressed in them. All studies should be English language full text publications that can be retrieved through electronic database or manual search.

Table 1

Operational definitions of domains, psychometric properties and subsets of psychometric properties for the purpose of this review

Exclusion criteria

If studies were aimed at the following research questions/designs, they will be excluded: (1) validating self-reported knee function and/or physical activity levels without addressing specific knee proprioception tests; (2) validating proprioception-related function23 such as dynamic balance, tendon tap (proprioceptive reflex), perturbation of actively positioned knee joint (proprioceptive reflex) or other nonspecific proprioception assessment methods such as a subjective questionnaire,34 a scale35 or other methods (cf. Röijezon et al. 2015)36; (3) validating measurement tools not specifically designed to assess knee joint proprioception (e.g. the Rivermead assessment of somatosensory performance)37; (4) investigating treatment effects following an intervention with a knee proprioception test without addressing PMPs of these tests; (5) pilot studies, abstracts, systematic reviews and meta-analyses, narrative reviews, book reviews, case series/reports, commentaries, editorials, letters to the editor, patient education handouts, consensus statements, clinical practice guidelines, theses/dissertations or unpublished literature; (6) non-English studies.

Information sources

One reviewer (AS) will conduct a systematic search in the following electronic databases: PubMed, AMED (via EBSCO), the Cumulative Index to Nursing and Allied Health Literature ((CINAHL) via EBSCO), SPORTDiscus (via EBSCO), Web of Science, Scopus, the Cochrane Central Register of Controlled Trials (CENTRAL), and Physical Education Index (via ProQuest).

Search strategy

The search strategy has been designed for a comprehensive search to locate the widest spectrum of articles for consideration using a combination of four conceptual groups of terms: (1) construct, (2) target population, (3) measurement instrument and (4) psychometric properties. In addition, an exclusion filter to omit secondary studies or publication types irrelevant to the current review will be applied. Depending on the electronic database searched, the search terms will be either keywords or database-specific search terms (MeSH, subject terms, subject headings and CINAHL headings) in combination with keywords or text words. Boolean operators, ‘OR’ and ‘AND’, will be used to combine the search terms. The search limits will be full text articles written in English language published from the inception of databases to November 2018. Titles and abstracts retrieved by the electronic search will be exported to EndNote library and duplicates will be excluded (AS). Screening of articles will be done using the following steps: the titles will be examined for relevance, then abstracts will be screened and finally the full text articles will be retrieved for data extraction and risk of bias assessment. Furthermore, any articles retrieved by hand search will be assessed for inclusion in a similar manner. The search strategy for each database is summarised in table 2. The article screening process will be depicted using the PRISMA flow chart.

Table 2

Search terms and strategy used for the electronic databases included in our systematic review

Study selection

Two reviewers (AS and ET) will independently evaluate titles and abstracts (and full text in case of doubt) of the retrieved references for eligibility using a screening questionnaire (box 1). Another reviewer (AA) will be consulted in case of any doubt in order to reach a consensus in determining the eligibility of studies to be included.

Box 1

A questionnaire to screen eligible studies for use at the title, abstract and full-text screening stages

Questions for all stages: title, abstract and full-text screening (follow steps 1 to 9)

  1. Is the study published in a scientific journal with a peer-review process?

    1. No—exclude

    2. Yes or uncertain—go to step 2

  2. Is the study published in English?

    1. No—exclude

    2. Yes or uncertain—go to step 3

  3. Does the study deal with adolescents (≥10 years) and/or adults?

    1. No—exclude

    2. Yes or uncertain—go to step 4

  4. Does this study investigate adolescents and/or adults with ACL injury (with or without reconstruction) and/or healthy individuals?

    1. No—exclude

    2. Yes or uncertain—go to step 5

  5. Does the article represent a primary study other than a case report/series (and not other types of research such as theses/dissertations, nonpeer-reviewed articles, letters to the editor, systematic reviews and meta-analyses, narrative reviews, book reviews, pilot studies, published study designs/trial protocols, commentaries, editorials, interviews, newspaper articles, patient education handouts, consensus statements or clinical practice guidelines)?

    1. No—exclude

    2. Yes or uncertain—go to step 6

  6. Does the study assess the knee?

    1. No—exclude

    2. Yes or uncertain—go to step 7

  7. Does the study employ any specific test to measure knee proprioception (e.g. JPS (active or passive; ipsilateral or contralateral matching), AMEDA, TTDPM, motion direction discrimination, pursuit/tracking task, force steadiness, force reproduction, velocity replication, velocity threshold hunting, psychophysical threshold hunting or any other related tests)?

    1. No—exclude

    2. Yes or uncertain—go to step 8

  8. Does this study report (objective) focal measures of knee proprioception (see point 7)?

    1. No—exclude

    2. Yes or uncertain— go to step 9

  9. Is the study designed to evaluate one or more measurement properties (validity, reliability, and/or responsiveness) of instruments/test procedures measuring knee proprioception?

    1. No—exclude

    2. Yes or uncertain—choose one of the following options:

      1. If ‘yes’—include at title and abstract screening stage

      2. If ‘uncertain’—follow steps 10–11

Additional questions for full-text stage only

  1. Does this study use at least one (appropriate) statistical test to analyse a psychometric property of a knee joint proprioception test’?

    1. No—exclude

    2. Yes or uncertain—go to step 11

  2. Are the points 1–10 scored as ‘yes’ or ‘uncertain’

    1. If all ‘yes’—include

    2. If any ‘uncertain’—discuss with another reviewer to come to an agreement whether to include the study or not

  • ACL, anterior cruciate ligament; AMEDA, active movement extent discrimination assessment; JPS, joint position sense; TTDPM, threshold to detect passive motion.

Data extraction process

The eligible articles will be divided between four reviewers (AA, AS, ET and UR) and a minimum of two reviewers (among AA, AS, ET and UR) will extract/verify the following data from each included study: (1) aims or research questions, (2) study design, (3) PMPs, (4) participant characteristics, (5) details of assessment (type of test, start position (e.g. sitting, standing, lying), equipment used, active or passive knee movement, velocity of knee motion, direction of motion, joint angle, muscle force, raters, number of sessions, time interval between trials/sessions, etc.), (6) objective outcome measures of knee joint proprioception including, among others, errors in joint position/angle matching (joint position sense error: constant error, variable error and absolute error), verbal identification of joint position,23 identifying location of joint motion,23 (passive) motion and direction detection,21 active reproduction of target force,19 21 weight identification,23 active reproduction of target velocity,21 the ability to discriminate between two or more movements,24 (7) data analysis specific to PMPs and (8) study findings (on validity, reliability, and responsiveness) and conclusions.

Risk of bias assessment of individual studies

A minimum of two reviewers (among AA, AS, ET and UR) will independently assess risk of bias of each included study using the updated COSMIN risk of bias checklist.38Any disagreement will be discussed and resolved by consensus and, if no consensus can be achieved, another reviewer (CH) will be available for cross-referral.

The COSMIN checklist is a standardised tool used to assess the risk of bias of studies investigating patient reported health outcomes. Nevertheless, in recent times, it has been used to assess the risk of bias of studies reporting measurement properties of physical performance tests39–41 using a four-point scoring system (very good, adequate, doubtful and inadequate). The criteria listed for reliability, measurement error, criterion validity, hypotheses testing for construct validity and responsiveness in the updated COSMIN checklist38 42 will be used for studies evaluating the PMPs of knee proprioception tests. The scoring will be done using the worst score of any standard in the box depicting the overall score in each subsection.

Planned methods of analysis

A qualitative narrative synthesis (text and/or table format) will include, but not limited to, sample size, participant characteristics, study design, specific knee proprioception tests, type of proprioception sense, direction of knee movement, knee joint angles, type of equipment, type of outcome measures, between-limb/group comparisons, data analysis and PMPs. In addition, risk of bias assessment scores achieved with the updated COSMIN tool will be summarised.

A meta-analysis using a random-effects model will be attempted when a minimum of three studies43 with relevant data on each knee proprioception test and with adequate homogeneity is available. Similar to papers44 45 published previously, a pooled estimate of appropriate statistical measures (e.g. intraclass correlation coefficient, area under the receiver operator characteristic curve, effect sizes, etc.) and their 95% CIs will be derived with forest plots along with estimates of statistical heterogeneity, wherever plausible. For instance, separate meta-analysis will be attempted to investigate inter-rater and intra-rater reliability; intraclass correlation coefficients from each study will be transformed to Fisher’s z scale and then pooled using a random-effects model.46 47 The weighted average values will then be converted back to intraclass correlation coefficient values again to allow interpretation of the findings.46 47 In cases where the same proprioception sense has been measured in different units in various studies then standardised mean difference instead of raw mean differences will be used for pooling the data in meta-analysis, if appropriate, for evaluating responsiveness. The standardised mean difference is the mean difference in outcome between groups or posttest vs pretest scores divided by the SD of outcome among participants.48 49 With the I2 statistic as a measure of statistical heterogeneity, a rough estimate of an I2 > 40% will be considered as a threshold for heterogeneity.49 At the same time, as the clinical relevance of heterogeneity present across studies (between-study variability) is important, τ2 statistic might be taken into account while performing meta-analysis.50

The feasibility of subgroup meta-analysis will be assessed based on the following factors: (1) each type of knee joint proprioception test and (2) population studied (healthy or those with ACL injuries managed conservatively or surgically). Moreover, if required, sensitivity analysis (the findings of high vs. low risk of bias studies) will be attempted in case of heterogeneous results. The trim and fill analyses (funnel plots) will be used to identify publication bias51 if at least 10 studies are included in the meta-analysis.52

When a meta-analysis is precluded then each PMP of knee proprioception tests will be evaluated using the criteria: sufficient (+), indeterminate (?) or insufficient (-) rating (table 3).53 Furthermore, the level of evidence for each knee proprioception test, based on the risk of bias scores of the included studies and the quality of PMPs, will be ascertained using an established criteria54 summarised in table 4.

Table 3

Criteria for evaluating the quality of the psychometric properties

Table 4

Levels of evidence rating for the quality of the psychometric properties

Ethics and dissemination

A systematic review of the available literature does not need ethical approval. Once the review is complete, it will be submitted as a series of manuscripts to scientific journals for peer-review and consideration for publication. The review findings may be presented at local and/or international conferences.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
View Abstract


  • Contributors AA and CKH conceived the idea of the project. AA was responsible for designing the review, conceptualising the initial review protocol and led the writing of the manuscript. AS, ET, UR and CKH contributed to the design of the review and drafting the manuscript. All authors have reviewed and revised the manuscript for important content and approved the final version of it. AA is the guarantor of this work.

  • Funding This review is conducted as a part of a larger ongoing project related to functional assessment of individuals with anterior cruciate ligament reconstruction, knee-healthy controls and elite athletes. Funding was received from the Swedish Scientific Research Council (Grant No. K2014-99X-21876-04-4), Västerbotten County Council (Grant No. ALF VLL548501), Swedish Scientific Research Council for Sports Science (Grant No. CIF 2016/6 P2017-0068). The funders were not involved in any aspect of the review protocol.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Patient consent for publication Not required.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.