Knee musculoskeletal (MSK) conditions significantly affect activities of daily living and physical performance in both the athletic and non-athletic populations.1-3 The assessment of physical function is key for the management of these conditions.4,5 Effective physical function measurement is important for assessing the burden of disease in a specific population, assessing the efficacy of different interventions in research, and tracking clinical intervention effectiveness and patient progress.6 Several measures of lower limb physical function have been developed and are available for people with knee MSK conditions, but many have poor psychometric properties, limited applicability for key populations (e.g., early knee osteoarthritis [OA]), and a narrow scope of assessment (i.e., performance tests that assess a single activity).5,7-9

Physical function, defined by the International Classification of Functioning, Disability and Health (ICF) as “the ability to move around and perform daily activities”, is complex to evaluate. Physical function represents multiple constructs such as different components (e.g. joint mobility, strength, balance, power), activity capacity (e.g. walking, climbing steps, kneeling, squatting) and ability to participate (e.g. employment, sports, hiking, gardening, playing with grandchildren).10 There is no gold standard for the assessment of lower limb physical function.5 There is a wide variety of self-reported outcome measures (SROMs) and performance-based outcome measures (PBOMs).11-13 SROMs and PBOMs measure different aspects of physical function, each with advantages and disadvantages.14,15 SROMs assess respondents’ perceptions of ability or limitation, whereas PBOMs objectively assess physical performance.16 SROMs tend to be less time-consuming to administer, can usually be completed independently by respondents, and are useful for conditions with fluctuating symptoms allowing the assessment of longer periods of time and average performance. However, SROM respondents may overestimate or underestimate their functional capacity, reducing the accuracy of the information collected.17 Self-reported data can also be affected by expectation, social desirability, and recall bias.18 These biases reduce the strength of evidence generated about interventions assessed by SROMs.19 In contrast, PBOMs have fewer biases (no recall bias or estimation bias, but may have measurement errors) and are generally more reproducible and reliable evaluations of activities of daily living and physical performance.20 In addition, PBOMs have greater sensitivity to change and less vulnerability to external influences such as cognition, culture, language, and education.21 The limitations of PBOMs include the assessment of a limited number of physical functions, the necessity of a competent observer for assessing and scoring, and higher time requirements than SROMs.22-24

Published systematic reviews of knee functional measures have studied the quality of the development process and the psychometric properties in specific populations (e.g., performance tests in people with anterior cruciate ligament [ACL] injuries).25-27 However, to our knowledge, no study has explored the content of these measures (i.e., which or how many physical functions are included). Identifying physical functions common to different measures will allow a better understanding of the spectrum of physical functions assessed and will help to identify gaps to address when developing new measures.

We aimed to explore the spectrum of physical functions assessed and how often they are assessed in existing SROMs and PBOMs used to measure knee function and physical performance in people with knee MSK conditions. Our secondary objectives were to compare the range and frequency of physical functions assessed in 1) different populations, and 2) different types of measure (PBOMs and SROMs).

We conducted a scoping review of the literature. Scoping reviews map key concepts underpinning a particular area of research and explore broad areas to identify gaps in the evidence, clarify concepts, and inform of the type of evidence particular to a research topic. A systematic review is well suited to address more precisely focused questions like assessing the psychometric properties of a specific measurement instrument but less suited to summarising a heterogeneous body of knowledge or identifying gaps in existing literature.28 This scoping review followed The Joanna Briggs Institute (JBI) Reviewers’ Manual 2015 and is reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses–Scoping Review (PRISMA-ScR) guidelines.28,29 The protocol is published in the Open Science Framework.30

Ordinal variables were summarized using counts and percentages. Continuous data were reported as mean and standard deviation (SD) or median and interquartile range (IQR), as appropriate. Rankings of the most assessed physical functions for each of the included populations were developed. We linked each extracted physical function to an ICF code.10 The ICF framework provides a standard language and a conceptual basis for the definition and measurement of health and disability. It is composed of 4 concepts: 1) body functions, 2) body structures, 3) activity and participation, and 4) environmental factors.33 Each concept can be subclassified in different subcategories. The linking steps were based on work from Cieza et al.36 linking rules with slight modifications. Firstly, the primary investigator (AP) established the link between the extracted physical function and the ICF code. In a second step, three investigators (BD, WT, and RS) reviewed this coding and reached consensus. Codes on which no consensus was reached were discussed until agreement was reached. For those physical functions allocated in “other, specified” ICF codes (e.g., d4108 “Changing basic body position, other specified”) inductive thematic analysis was used to create subgroups within these codes. Measurement properties were extracted and grouped according to the COSMIN taxonomy into the four categories: validity (content validity, construct validity, and criterion validity), reliability (test-retest reliability, internal consistency, and measurement error), responsiveness and interpretability (floor and ceiling effects, minimal detectable change, minimum clinical important difference, and substantial clinical benefit).32

After removal of the duplicates, we screened 4173 titles and abstracts and reviewed 215 full-text articles (186 identified from databases and 29 from citation searching). After excluding 72 articles, data were extracted from 143 articles (Fig. 1). Eighty-eight studies included participants with knee OA, 49 included participants with ACL injuries, and 19 included participants with PFP (7 studies included two populations and 3 studies included all three populations). Compared to ACL and PFP studies, knee OA studies included larger (median number of participants 93 vs. 73 and 50, respectively) and older (mean age 62 years old vs. 28 and 29) populations. Knee OA and PFP studies included a greater proportion of females (63% and 54%, respectively) than ACL studies (42%) (Supplementary material 2 – Tables A, B, and C). Reliability was the most reported measurement property in knee OA (81%) and PFP (73%) studies, whereas validity was the most reported in ACL studies (71%).

Fig. 1.

PRISMA flow diagram for the Scoping review.

(0.57MB).

710 physical functions were extracted from 93 different measures and mapped to 50 single ICF codes. The 15 most assessed physical functions are presented in Table 2. An additional 56 physical functions can be found in an extended table in the Supplementary material 3. These other 56 physical functions were assessed between 1 and 12 times, with 30 of these being assessed only once. Climbing stairs was the most assessed physical function in measures for knee OA and PFP, whereas jumping was the most assessed in measures for ACL injuries.

Of the 93 extracted measures, 53 were PBOMs, 37 were SROMs, and 3 were mixed measures (i.e., include both self-reported items and performance tests). Fifty-six measures assessed knee OA (31 PBOMs, 24 SROMs, and 1 mixed measure), 34 ACL injuries (18 PBOMs, 14 SROMs, and 2 mixed measures), and 19 PFP (7 PBOMs, 11 SROMs, and 1 mixed measure). Fig. 2 shows the number of measures for each or multiple populations.

Fig. 2.

Number of measures across the different populations. Each circle represents a specific knee population. The size of the circle corresponds to the number of measures for each population. Shared areas represent measures common for more than one population (i.e., knee OA has 56 measures in total). Abbreviations: ACL, anterior cruciate ligament; OA, osteoarthritis; PFP, patellofemoral pain.

(0.12MB).

This scoping review explored the content of measures used to evaluate lower limb physical function in people with knee MSK conditions. We identified 93 different measures that assessed 710 discrete physical functions that were linked to 50 ICF codes. The majority of the measures were developed/assessed in people with knee OA. SROMs assessed a much larger spectrum of physical functions than PBOMs. Climbing stairs and standing up from sitting were the most commonly assessed functions in SROMs, whereas walking short distances and jumping were the most commonly assessed functions in PBOMs. Measures for people with ACL injuries assessed more dynamic and challenging functions than those for people with knee OA and PFP, which focused in lower difficulty daily tasks.

Overall, the five most assessed physical functions were climbing stairs, standing up from sitting, walking short distances, jumping, and squatting. However, we found slight differences in the rankings within each clinical population. Weight-bearing activities (e.g., climbing stairs, standing up from sitting) were the most assessed physical activities in measures for people with knee OA and PFP. Jumping was the most assessed physical function in measures for people with ACL injuries. The most commonly assessed physical functions for each population reflect key symptoms and impairments (commonly impacted activities) or the reason for the test (such as assessing readiness to return to sport in people with ACL injuries).37-43 However, this focus on conditions’ key characteristics might limit assessment to a narrower level of function and reduce applicability to diverse populations or symptom states. Challenging and high-impact activities such as jumping and running were highly assessed within measures for people with ACL injuries (ranked 1st and 3rd, respectively), moderately assessed in measures for PFP (ranked 7th and 5th, respectively), but rarely assessed (ranked 20th and 13th, respectively) in measures for knee OA. This may be due to people with ACL injuries and PFP commonly being young and physically active. However, this is evidence of a gap in the assessment of demanding activities in people with knee OA. Assessing these demanding activities could be important for the detection of people with early symptoms of knee OA who might also be young and physically active or in need of monitoring functional change over time.38,44,45

There was a large contrast in the number of physical functions assessed by self-reported and performance-based measures. PBOMs assessed fewer physical functions than SROMs (median of 1 vs. 9). This means that multiple PBOMs may be necessary to have complete objective evaluation of a person's physical ability, however, this would be difficult to achieve given the different scoring methods used. Unilateral activities were more commonly assessed in PBOMs than in SROMs. This suggests that performance tests might be better suited to detect differences between injured and non-injured limbs.

Most physical functions could be linked to an ICF code. However, some issues were identified through this process. We identified several items in SROMs that attempted to measure multiple physical functions (e.g., “bending, kneeling, or stooping” or “vacuuming or yard work”). These multiple function items are likely to generate interpretation and response variability affecting measurement quality. Some items were vague in their description (e.g., “stairs”, “sit”, “standing”), similarly allowing broad interpretation. The ICF lacked codes for some activities that were frequently assessed in measures and represent common activities of daily living such as ‘getting in/out of vehicles’; ‘turning, twisting, or pivoting’ and ‘getting in/out of a bath’. Likewise, challenging activities such as ‘cuts/sharp turns’ or ‘accelerating/decelerating while running’ lacked specific ICF codes. Some of these issues have been reported previously and may represent an opportunity to improve the ICF.46,47

Key gaps we identified in available measures to assess knee physical function are the inclusion of very few physical functions in PBOMs, knee OA measures generally not including challenging activities, and high variability among the wording and specificity of items in SROMs. Addressing these gaps when selecting or drafting items/activities can help clinicians and researchers develop better and more comprehensive measures. Based on the findings of this research, we suggest: 1) including several physical functions across different levels of difficulty that can be scored using the same method in a single PBOM to enable generation of a single score that reflects an individual's overall lower limb physical function; 2) avoiding use of multi-activity items in SROMs given these might be interpreted differently by different respondents; 3) developing SROMs that include unilateral functions; 4) including more challenging activities in scales for OA to better reflect the diverse ability in this group; and 5) being as specific as possible when drafting questionnaire items. Items that are vague or not detailed enough may be hard to interpret, and therefore, difficult to rate consistently. For example, there is a clear difference between “climbing stairs” and “climbing up and down 3 flights of stairs indoors, using a handrail”. Even though both items assess climbing stairs, the latter describes a more precise context to rate.

Measures to assess lower limb physical functions include a wide variety of activities. Climbing stairs is the most assessed physical function in outcome measures for knee OA and PFP populations, whereas jumping is the most assessed physical function in outcome measures for the ACL injury population. Challenging and more difficult activities are rarely assessed in the knee OA outcome measures. SROMs generally assess a broader range of bilateral physical functions, whereas PBOMs focus on discrete physical activities (bilateral or unilateral) and are difficult to combine due to different scoring systems. Current physical function outcome measures are not well suited to assess performance in OA populations with mild or diverse levels of impairment.

The review protocol, as well as the complete dataset are available on Open Science Framework, a public open-access repository.