You are currently on the new version of our website. Access the old version .
MDPIMDPI
Search all articles
DisabilitiesDisabilities
Download PDF
  • Study Protocol
  • Open Access

27 January 2026

Longitudinal Predictors of Pain and Physical Function Trajectories over 12 Months in Older Adults with Knee Osteoarthritis Receiving an Education and Exercise Program: Statistical Analysis Protocol

,
and
1
Faculty of Health Sciences, University of Málaga, 29010 Málaga, Spain
2
Department of Physiotherapy, University of Valencia, 46010 Valencia, Spain
3
IIMPACT in Health, Allied Health and Human Performance, University of South Australia, Adelaide 5001, Australia
*
Author to whom correspondence should be addressed.
Disabilities2026, 6(1), 14;https://doi.org/10.3390/disabilities6010014
Version Notes
Order Reprints

Abstract

Knee osteoarthritis (KOA) is a leading cause of disability in older adults, characterized by persistent pain and reduced physical function. Beyond localized joint pathology, many individuals with knee osteoarthritis experience multisite pain and live with multiple comorbidities, reflecting a heterogeneous and multifactorial pain condition. Prognostic models based primarily on biomedical variables have shown limited ability to explain long-term outcomes, partly due to insufficient integration of pain chronicity, comorbidity count and psychosocial determinants such as treatment expectations and pain self-efficacy. While exercise and education are commonly recommended as primary non-surgical treatments, people often respond to them very differently. This study protocol describes a secondary longitudinal observational analysis of data from the EPIPHA-KNEE two-arm, multicentre randomized controlled trial. The primary outcomes will be knee OA pain intensity and physical function, assessed using the Western Ontario and McMaster Universities Arthritis Index (WOMAC) questionnaire at baseline, 3, 6 and 12 months. Baseline prognostic factors will include pain duration, pain distribution, comorbidity count and patient expectations, including treatment expectations and pain self-efficacy. Linear mixed-effects models will be used to examine longitudinal associations between these predictors and pain and function trajectories, with particular emphasis on predictor-by-time interactions to characterize differential patterns of change over time. The planned analyses aim to improve understanding of how clinical characteristics and expectancy-related factors jointly shape 12-month pain and physical function trajectories in older adults with knee osteoarthritis receiving education and exercise-based care, thereby informing prognostic stratification within non-surgical management.

1. Introduction

Knee osteoarthritis (KOA) is one of the most common musculoskeletal disorders and a leading cause of disability worldwide, particularly in older adults. It manifests with pain, stiffness, and functional limitations that severely compromise quality of life and independence [1]. While traditionally considered a localized joint disease, KOA is increasingly recognized as part of a broader, multifactorial pain syndrome, in which symptoms cannot be explained by structural joint damage alone [2,3]. Accumulating evidence indicates that KOA-related pain arises from the interaction of peripheral nociceptive input from joint tissues, inflammatory and biomechanical factors and multisite pain involvement [4,5]. In addition, psychosocial and contextual factors substantially influence pain experience, functional limitation, and response to treatment [6]. Many patients experience pain beyond the knee, including the hips, spine, and contralateral joints, suggesting potential involvement of widespread pain mechanisms [7,8]. These insights challenge the conventional biomedical model of KOA, which often fails to adequately predict patient trajectories of pain and disability [9].
Large-scale prognostic studies have attempted to develop predictive models based on biomedical and demographic factors, but their ability to anticipate individual patient outcomes has been limited [10]. This limitation underscores the complexity of KOA, in which biological, psychological, and social factors interact. Widespreadness of pain is increasingly recognized as an important factor in KOA, adding complexity to treatment planning and management [11].
Multimorbidity further amplifies this complexity. Most older adults with symptomatic knee osteoarthritis also present with at least one additional chronic condition, such as obesity, diabetes, cardiovascular disease, or depression [12,13]. These comorbidities not only worsen KOA-related symptoms but also reduce physical activity and increase healthcare utilization [14]. Importantly, multimorbidity often leads to polypharmacy, as patients rely on multiple pharmacological treatments to manage both KOA and coexisting conditions [15]. While pharmacological therapies such as non-steroidal anti-inflammatory drugs (NSAIDs) and intra-articular corticosteroids remain cornerstones of symptomatic treatment, they are associated with adverse effects, particularly in older populations with fragile health [16]. The use of opioids, which are sometimes prescribed for refractory KOA pain (i.e., ongoing knee pain that persists despite appropriate exposure to guideline-recommended first-line non-surgical care, including structured exercise and education [17] is particularly concerning due to risks of dependency, overdose, and poor long-term efficacy [18,19]. Recent reviews highlight that inappropriate opioid prescribing remains a persistent issue in KOA and other chronic musculoskeletal conditions, especially in patients with comorbidities and high psychological distress [20,21].
High healthcare utilization is a common consequence of this complex clinical picture. Patients with knee osteoarthritis and multimorbidity have greater reliance on specialist care, emergency visits, and hospitalizations, often related to pain exacerbations or complications of comorbid diseases [22]. This increased healthcare burden is often accompanied by greater reliance on passive treatment strategies, such as manual therapy, massage and electrotherapy, that involve minimal “active” patient participation [23]. Use of passive treatment strategies may further discourage patient engagement in future active rehabilitation programs [24]. Such disengagement is problematic because non-pharmacological approaches, particularly exercise therapy, remain the cornerstone of KOA management [25]. Exercise and education are associated with improvements in pain, physical function and quality of life, with recent evidence showing that the magnitude and sustainability of these benefits are strongly influenced by patient adherence and engagement, contributing to substantial inter-individual variability in outcomes [26,27].
Psychosocial factors, especially expectations and self-efficacy, are increasingly recognized as determinants of treatment outcomes in KOA. Patient expectations influence not only perceived pain and functional improvement but also engagement and adherence to exercise [28,29]. Positive expectations are associated with reduced pain and higher levels of physical activity, whereas negative expectations predict poor adherence and worse outcomes [30]. Importantly, a large cohort study demonstrated that high self-efficacy for managing KOA symptoms predicted reduced pain and greater physical activity at both 3- and 12-month follow-up [31]. Conversely, other comorbidities appear to attenuate the protective effects of self-efficacy, highlighting the need to contextualize these psychosocial variables within complex clinical presentations [32].
Despite this evidence, measuring expectations remains inconsistent. Current studies employ heterogeneous tools, ranging from single-item questions to elaborate multidimensional scales, limiting comparability across trials and systematic reviews [26]. Moreover, expectations are rarely integrated into prognostic models alongside biological and comorbidity-related factors, which constrains our understanding of how they interact with other prognostic domains. This methodological gap reduces the ability to identify subgroups of patients who may benefit most from expectation-focused interventions or stratified care approaches [27].
Complex KOA cases, characterized by widespread pain and multiple comorbidities, underscore the importance of realistic expectation-setting. Both patients and clinicians must recognize at treatment onset that complete recovery is unlikely in many cases, and that meaningful improvements may involve partial but clinically significant gains [33]. If expectations are overly optimistic and not adjusted to the patient’s clinical profile, treatment failure may be perceived even when objective improvements are achieved, undermining adherence and satisfaction [34,35]. Conversely, interventions that enhance self-efficacy and align expectations with realistic outcomes have the potential to improve engagement and long-term trajectories [36].
Thus, KOA should be conceptualized not merely as a local joint disease but as a complex chronic condition at the intersection of nociceptive, systemic, and psychosocial processes. Effective prognostic models and management strategies must therefore account for multimorbidity and psychosocial factors, including expectations and self-efficacy. Yet, significant gaps remain: (1) the duration and extension or widespreadness of pain are insufficiently integrated into prognostic tools; (2) comorbidities are often treated as confounders rather than key modifiers of outcomes; and (3) expectations, though consistently linked to clinical trajectories, are rarely contextualized within complex patient phenotypes. Addressing these gaps will allow the development of stratified care pathways that align therapeutic approaches to patient risk profiles and needs.
Therefore, this study will explore the prognostic value of pain duration, pain extent, comorbidities, and expectations in predicting response to treatment trajectories in older adults with knee osteoarthritis, with the ultimate aim of informing stratified, patient-centred care.

1.1. Objectives

1.1.1. Primary Aim

To longitudinally investigate how patient expectations (including treatment expectations and self-efficacy), baseline pain duration, pain distribution and comorbidity count influence 12-month trajectories of pain and physical function in older adults with knee osteoarthritis receiving an education and exercise program. This aim will enable the identification of predictors of treatment responses and longitudinal treatment trajectories within non-surgical care.

1.1.2. Secondary Aims

To describe longitudinal changes in pain and physical function over 12 months, and to visualize differential trajectories across groups defined by baseline values (pain duration, pain distribution, comorbidity count, treatment expectations, and self-efficacy) in older adults with knee osteoarthritis receiving an education and exercise program. This aim will facilitate the interpretation of predictor-by-time interactions and illustrate differential symptom evolution for clinical context.
We hypothesize that longer pain duration, more widespread pain, higher comorbidity count, negative treatment expectations, and lower baseline self-efficacy will be significantly associated with higher pain intensity and lower physical function trajectories over 12 months.

2. Materials and Methods

2.1. Study Design

Secondary analysis of data from the EPIPHA-KNEE project, a two-arm, multicentre RCT, prospectively registered (ACTRN12620001041943; https://bit.ly/2SfVySS, accessed on 13 October 2020) and approved by the University of South Australia Human Research Ethics Committee (HREC No. 20237), the Central Adelaide Local Health Network Human Research Ethics Committee (HREC No. 12579), the University of Melbourne Human Research Ethics Committee (HREC No. 2057540), and Flinders University Human Research Ethics Committee (HREC No. 4478). The present study will treat the dataset as a longitudinal observational cohort study. The Declaration of Helsinki, the Guidelines for Good Clinical Practice (GCP) and the Strengthening of the Reporting of Observational Studies in Epidemiology (STROBE) statement for observational studies will be followed. The Prognostic Research Strategy (PROGRESS) framework will be used to incorporate best practices for prognostic research as part of this study protocol.

2.2. Participants and Setting

Full recruitment information and data collection have already been published [37].
198 patients aged ≥ 50 years with knee osteoarthritis according to the National Institute for Health and Care Excellence (NICE) clinical criteria, with pain for at least 6 months, who have at least moderate levels of pain (≥4 on an 11-point Numerical Pain Rating Scale) and who report at least moderate difficulty with daily activities. NICE criteria for age is ≥45 years; however, to limit inclusion of younger participants with trauma-induced osteoarthritis, ≥50 years have been used for the present study.
Participants were randomly assigned to either best practice care, involving standard education and an individualised walking and lower limb strengthening program, or to pain science-informed care, involving pain science education and an individualised walking and lower limb strengthening program. Both intervention arms include a consistent, standardized approach to general osteoarthritis (OA) and physical activity education, incorporating the structured walking and strengthening exercises. However, they differ in the content of their OA pain education. The EPIPHA-KNEE group received a contemporary pain science education (PSE) program, grounded in self-regulated learning strategies and conceptual change theory, and adapted to incorporate up-to-date biological insights into OA. In contrast, the Best Practice Care Control group received education aligned with current clinical guidelines for OA management. Both groups received 4 weeks of weekly in-person sessions with a physiotherapist, followed by 4 weeks of weekly telehealth sessions, a telehealth session at 12 weeks, and in-person sessions at month 5 and 9. A detailed description of the exercise protocols and a comparative table outlining the educational content of both intervention arms have been published previously [37].

2.3. Sample Size

Based on commonly accepted recommendations for regression modelling in observational studies, a minimum of 10 participants per independent variable was considered necessary to ensure reliable and stable estimates, avoiding overfitting. Based on the inclusion of up to 15 predictors in the planned multivariable models, a minimum sample size of 150 participants was considered necessary. To account for an estimated 20% attrition or missing data over the one-year follow-up period, the final target sample size was increased to at least 180 participants.

2.4. Measures and Procedures

2.4.1. Baseline Variables

At baseline, participants completed sociodemographic data (age, gender, height and weight, body mass index, postcode, and educational level), current/recent treatment exposure (i.e., pharmacological, surgical, conservative [e.g., physiotherapy], complementary [e.g., acupuncture]), presence of comorbidities, expectations about treatment and their ability to cope with pain or pain self-efficacy and pain history (duration of symptoms and pain extension).

2.4.2. Dependent Variables of Interest

The dependent variables in the present study include:
OA knee symptoms: pain intensity and physical function were measured by the Western Ontario and McMaster Universities Arthritis Index (WOMAC) questionnaire pain and physical function subscales. The WOMAC is a disease-specific self-report questionnaire that includes pain (5 items), stiffness (2 items), and physical function (17 items) subscales. Items are scored from 0 to 4, giving a range of possible scores from 0 (no symptoms or dysfunction) to 96 (maximal symptoms and dysfunction). It has demonstrated validity, reliability, and responsiveness. The primary time points for both outcomes are 3, 6 and 12 months.

2.4.3. Independent Variables

The independent variables for the present study include the following variables assessed at baseline:
(i)
Pain extent assessed by selecting items in a list of problems in other joints. The instructions stated: “Do you suffer from any problems (e.g., pain, aching, discomfort or stiffness) around the following joints in your body? Put a cross in as many as apply to you”. The responses were classified into four groups adapted from a suggested new classification system for low back pain. (a) Strict chronic local pain: pain in one joint (knee), no further pain. (b) Chronic regional pain: pain in one knee plus additional joint pain, but criteria for chronic widespread pain (CWP) is not fulfilled. (c) Common CWP: pain in at least two contra-lateral joints (one upper and one lower), but less than four limbs. (d) Extreme CWP: pain in all four limbs joints and spinal areas. Patients with pain in other body areas that currently limits walking ability were excluded to ensure that knee-related functional limitations were the primary contributor to the outcomes analysed.
(ii)
Comorbidity count via the Functional Comorbidity Index, an 18-item list of diagnoses, each of which is given 1 point if present, and the final score is the sum of the items, indicating the number of comorbidities. This approach provides a pragmatic measure of overall comorbidity profile but does not differentiate comorbidities by duration, severity or treatment exposure.
(iii)
Expectations of treatment outcomes assessed using a 5-point ordinal scale in response to the question “What effect do you think this treatment will have on your knee problem”, with choices from “no effect at all” to “complete recovery”. A single-item measure of recovery expectations can validly be used in individuals with musculoskeletal pain conditions.
(iv)
Expectations about ability to cope with pain or pain self-efficacy measured by the Pain Self-efficacy Questionnaire (PSEQ), which includes 10 items scored on a 7-point Likert scale from “Not at all confident” (0) to “Completely confident” (6), with scores ranging from 0 (no confidence to perform activity despite pain) to 60 (maximal confidence to perform activity despite pain).

2.5. Statistical Analysis

All statistical analyses will be conducted in R (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria) using the RStudio environment (version 4.5). Mixed models will be fitted using lme4, multiple imputation will be conducted with mice, correlations will be analysed using psych, and graphics will be produced with ggplot2. Full analysis code will be archived for reproducibility.

2.5.1. Descriptive Analysis

Descriptive statistics will be used to characterize the sample at baseline. Continuous variables will be summarized as means with standard deviations (SD) and 95% confidence intervals (CI), and categorical variables will be summarized as absolute and relative frequencies. Baseline sociodemographic, physical and psychosocial characteristics will be compared using independent t-tests for continuous variables and chi-square tests for categorical variables, as appropriate, to ensure comparability between treatment groups. Standardized mean differences will be calculated to facilitate comparison across baseline characteristics. These exploratory analyses will also be used to examine potential attrition patterns at 12 months by comparing baseline characteristics between participants with and without complete follow-up.

2.5.2. Correlation Analysis

Prior to longitudinal modelling, correlations between baseline predictors (pain duration, pain distribution, comorbidity count, treatment expectations and pain self-efficacy) and between these predictors and baseline WOMAC pain and physical function will be examined using Pearson’s correlation coefficient. Correlation strength will be interpreted as strong (>0.60), moderate (0.30–0.60) or weak (<0.30). These analyses will inform the assessment of multicollinearity and conceptual overlap before multivariable modelling. Correlation matrices will be visualized using heatmaps to enhance interpretability.

2.5.3. Longitudinal Mixed-Effects Modelling

Longitudinal changes in WOMAC pain and physical function from baseline to 3, 6 and 12 months will be analysed using linear mixed-effects models (LMM) fitted with the lmer() function from the lme4 package (version 1.1–38). LMMs will be selected because they account for within-person correlation across repeated measures, allow inclusion of participants with incomplete follow-up under a Missing At Random (MAR) assumption, and provide more efficient estimates than complete-case approaches.
Random intercepts will be specified for participants, and random slopes for time will be included when supported by model fit. Time will be treated as a categorical variable to allow for non-linear changes in outcomes across measurement occasions. Fixed effects will include time, each baseline predictor, and predictor × time interaction terms. Predictors will include pain duration, pain distribution (local/regional vs. generalized/highly generalized or number of painful regions), comorbidity count, treatment expectations and pain self-efficacy. The predictor × time interaction will be the primary parameter of interest, indicating whether baseline prognostic factors are associated with differential trajectories of pain or physical function over 12 months. We will also use a tripartite variance analysis (i.e., identifying random variance, variance associated with the predictors, and variance associated with extrinsic/unknown variable[s]) to explore the proportion of unexplained variance associated with systemic effects not included in the model (e.g., associated with treatment group allocation of the original RCT) [38,39].
A fully adjusted multivariable model will then be fitted including all baseline predictors and a priori covariates/confounders (age, sex, body mass index, baseline prior treatment exposure, baseline WOMAC pain or physical function, and other relevant clinical variables). A similar tripartite variance analysis will be completed to explore the proportion of unexplained variance associated with systemic effects not included in the model. Model assumptions (linearity, normality of residuals and homoscedasticity) will be evaluated using graphical diagnostics and residual statistics. If assumption violations are detected, alternative model specifications (including transformations or spline terms for non-linear effects) will be explored. Effect sizes will be reported alongside p-values. Partial eta squared (ηp2) will be calculated for fixed effects and interpreted as small (≥0.01), medium (≥0.06) or large (≥0.14). When significant overall time effects are observed, post hoc pairwise contrasts between time points will be conducted using Bonferroni-adjusted tests. Cohen’s d will be used to quantify the magnitude of change between successive time points (0.20–0.49 small, 0.50–0.79 medium, ≥0.80 large). Cohen’s d was used to quantify the magnitude of change between successive time points (0.20–0.49 small, 0.50–0.79 medium, ≥0.80 large).

2.5.4. Missing Data and Imputation

Patterns of missing data will be examined descriptively. Longitudinal mixed-effects models will be estimated using maximum likelihood, which provides unbiased estimates under the MAR assumption given the observed data. For analyses requiring complete baseline predictors, multiple imputation by chained equations (MICE) will be used to generate 20 imputed datasets, incorporating outcomes, predictors and auxiliary variables associated with missingness. Imputed datasets will be converted to long format and analysed using the same LMM structure, and estimates will be pooled using Rubin’s rules. Sensitivity analyses using complete cases will be performed to assess the robustness of findings.

2.5.5. Trajectory Visualization

To enhance interpretability, mean trajectories of WOMAC pain and physical function across time will be plotted with 95% confidence intervals. For predictors showing clinically or statistically relevant predictor × time interactions, additional stratified trajectory plots (e.g., high vs. low pain duration or high vs. low pain self-efficacy) will be generated to illustrate differential symptom evolution. This visualization strategy will support interpretation of prognostic subgroups without relying on class-based modelling.

3. Discussion and Perspectives

This study protocol describes a secondary longitudinal analysis of the EPIPHA-KNEE cohort aimed at examining how baseline clinical characteristics of pain duration, pain distribution and comorbidity count together with patients’ expectancy-related factors (treatment expectations and pain self-efficacy) are associated with trajectories of pain intensity and physical function over 12 months.
Although the present analysis treats the cohort as a longitudinal observational study rather than a comparative effectiveness trial, both groups received structured exercise and guideline-consistent education, providing a clinically relevant context in which to investigate prognostic factors within contemporary non-surgical care [23,25]. Further, use of tripartite variance analysis ensures consideration of systemic effects not included in the model, such that potential contribution of the original RCT treatment group assignment is formally explored.
Pain duration, multisite pain and comorbidity count are included as core prognostic variables based on consistent evidence linking chronicity, widespread pain and multimorbidity to poorer outcomes in KOA and other musculoskeletal pain conditions [12,32,40,41,42,43]. Pain distribution is operationalized using a pragmatic classification based on the number and location of painful regions, ensuring clinical relevance and feasibility while remaining aligned with the study objectives. The EPIPHA-KNEE exclusion criteria were applied to ensure that knee-related functional limitations were the primary contributor to disability, supporting internal consistency of the prognostic analyses.
Comorbidity count is quantified using the Functional Comorbidity Index, allowing assessment of cumulative disease load rather than isolated diagnoses. While the duration and severity of individual comorbidities are not modelled, this approach aligns with the prognostic focus of the study and avoids over-parameterization. Imaging-derived inflammatory features is recognized as clinically relevant in knee osteoarthritis; however, detailed information on imaging findings was not consistently available across the EPIPHA-KNEE cohort and therefore could not be incorporated into the present analytical framework. Accordingly, the study is designed to test predefined prognostic hypotheses using variables that are consistently captured and clinically scalable, rather than to model all possible determinants of pain [44].
Expectancy-related variables are included given their established influence on pain outcomes, adherence to exercise and engagement with self-management strategies [45,46,47]. The use of a single-item recovery expectation measure [48] and the Pain Self-Efficacy Questionnaire [49] reflects a deliberate balance between psychometric robustness and clinical feasibility, enhancing translational potential. The planned examination of predictor-by-time interactions will allow assessment of whether these psychosocial factors act as independent prognostic markers or modify longitudinal symptom trajectories associated with clinical risk factors.
The longitudinal mixed-effects modelling strategy accommodates repeated measures, accounts for within-person correlation and allows inclusion of participants with incomplete follow-up under a missing-at-random assumption. Treating time as a categorical variable permits non-linear symptom trajectories to be modelled, and stratified trajectory visualisation is included to enhance clinical interpretability beyond statistical significance alone.
As a secondary observational analysis, the study is not intended to establish causal relationships, and residual confounding cannot be excluded. Further, predictive results will be limited as a function of outcome assessment and available data. For example, pain in knee osteoarthritis can be assessed using a range of validated instruments, including WOMAC [50] pain, Knee injury and Osteoarthritis Outcome Score (KOOS) [51] pain visual analogue scales (VAS) and numerical rating scales (NRS), each capturing partially distinct aspects of the pain experience. In line with the study objectives and the design of the parent trial, WOMAC pain and physical function are used as the primary outcomes, given their widespread use and established validity in KOA research. While no single instrument fully captures the multidimensional nature of pain, the use of WOMAC pain provides a standardised and clinically meaningful measure for longitudinal prognostic analyses [23]). Similarly, comorbidity profile is limited to count data of the number of co-morbid conditions and individual reports, which precludes exploring the impact of the severity or duration of these co-morbidities on pain or functional outcome. Finally, eligibility criteria related to the trial, whereby those who had pain in other body areas that primarily limited walking ability were excluded, may result in a sample less severely affected by pain. As such, it is possible that associations between pain extent and long-term pain and function may be of smaller magnitude than seen in a more severely affected population. Despite these limitations, overall, this protocol aims to generate methodologically rigorous and clinically actionable prognostic evidence to support the development of risk-stratified, patient-centred non-surgical care pathways for knee osteoarthritis.
Lastly, by combining clinical phenotype (pain extent and duration), comorbidity profile and readily obtainable expectancy/self-efficacy measures within longitudinal models, this study aims to shift knee OA management from a one-size-fits-all paradigm toward a risk-stratified, patient-centred approach. Even in the presence of methodological constraints inherent to secondary analyses, robust and reproducible prognostic findings would offer a pragmatic foundation for targeted interventions and optimise resource allocation in ageing populations living with knee osteoarthritis.

Author Contributions

M.F.-C. conceived the idea for the study. T.R.S. led the trial. M.F.-C., F.C.-M. and T.R.S. designed the protocol. M.F.-C. drafted the manuscript with input from others. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Ethical approval for the EPIPHA-KNEEtrial was obtained from the University of South Australia Human Research Ethics Committee (HREC No. 20237, 12 September 2019), the Central Adelaide Local Health Network Human Research Ethics Committee (HREC No. 12579, 1 June 2020), the University of Melbourne Human Research Ethics Committee (HREC No. 2057540, 8 June 2020), and Flinders University (HREC No. 4478, 7 February 2020).

Data Availability Statement

The datasets used and analysed during the current study will be available from the corresponding author on reasonable request once the study has been completed.

Acknowledgments

Mar Flores-Cortés is a predoctoral student of the University of Malaga. We would like to thank Ty Stanford for his contribution of biostatistical expertise.

Conflicts of Interest

The authors declare no conflicts of interest.

Disability Language/Terminology Positionality Statement

In this manuscript, we primarily use person-first language (e.g., “people with knee osteoarthritis”, “older adults living with knee osteoarthritis”) because our work is situated in clinical rehabilitation and health research contexts where person-first terminology is commonly used and aligns with rights-based approaches. We use the term “disability” to refer to activity limitations and participation restrictions related to symptoms and functional impact, rather than as a label for individuals. We acknowledge that language preferences vary and that some people and communities prefer identity-first language (e.g., “disabled people”); where individual preferences are known (e.g., in direct quotations), we would use the terminology preferred by those individuals.

References

  1. Cross, M.; Smith, E.; Hoy, D.; Nolte, S.; Ackerman, I.; Fransen, M.; Bridgett, L.; Williams, S.; Guillemin, F.; Hill, C.L.; et al. The global burden of hip and knee osteoarthritis: Estimates from the Global Burden of Disease 2010 study. Ann. Rheum. Dis. 2014, 73, 1323–1330. [Google Scholar] [CrossRef] [PubMed]
  2. Hunter, D.J.; Bierma-Zeinstra, S. Osteoarthritis. Lancet 2019, 393, 1745–1759. [Google Scholar] [CrossRef] [PubMed]
  3. Neogi, T. The epidemiology and impact of pain in osteoarthritis. Osteoarthr. Cartil. 2013, 21, 1145–1153. [Google Scholar] [CrossRef] [PubMed]
  4. Korochina, K.V.; Chernysheva, T.V.; Korochina, I.E. The Impact Of Comorbid Conditions On Clinical Manifestations Of Knee Osteoarthritis Grades 3-4. Osteoarthr. Cartil. 2023, 31, S145. [Google Scholar] [CrossRef]
  5. Girbés, E.L.; Nijs, J.; Torres-Cueco, R.; Cubas, C.L. Pain treatment for patients with osteoarthritis and central sensitization. Phys. Ther. 2013, 93, 842–851. [Google Scholar] [CrossRef]
  6. Moreira, V.M.P.S.; Signorelli, F.S.; Hattori, W.T.; Dionisio, V.C. Association between psychological factors and physical performance in individuals with knee osteoarthritis: A cross-sectional study. J. Bodyw. Mov. Ther. 2025, 42, 274–282. [Google Scholar] [CrossRef]
  7. Tang, S.; Zhang, C.; Oo, W.M.; Fu, K.; Risberg, M.A.; Bierma-Zeinstra, S.M.; Neogi, T.; Atukorala, I.; Malfait, A.-M.; Ding, C.; et al. Osteoarthritis. Nat. Rev. Dis. Prim. 2025, 11, 10. [Google Scholar] [CrossRef]
  8. Arendt-Nielsen, L.; Nie, H.; Laursen, M.B.; Laursen, B.S.; Madeleine, P.; Simonsen, O.H.; Graven-Nielsen, T. Sensitization in patients with painful knee osteoarthritis. Pain 2010, 149, 573–581. [Google Scholar] [CrossRef]
  9. Baumbach, L.; List, M.; Grønne, D.T.; Skou, S.T.; Roos, E.M. Individualized predictions of changes in knee pain, quality of life and walking speed following patient education and exercise therapy in patients with knee osteoarthritis—A prognostic model study. Osteoarthr. Cartil. 2020, 28, 1191–1201. [Google Scholar] [CrossRef]
  10. Bastick, A.N.; Belo, J.N.; Runhaar, J.; Bierma-Zeinstra, S.M.A. What Are the Prognostic Factors for Radiographic Progression of Knee Osteoarthritis? A Meta-analysis. Clin. Orthop. Relat. Res. 2015, 473, 2969–2989. [Google Scholar] [CrossRef]
  11. Dell’ISola, A.; Pihl, K.; Turkiewicz, A.; Hughes, V.; Zhang, W.; Bierma-Zeinstra, S.; Prieto-Alhambra, D.; Englund, M. Risk of Comorbidities Following Physician-Diagnosed Knee or Hip Osteoarthritis: A Register-Based Cohort Study. Arthritis Care Res. 2022, 74, 1689–1695. [Google Scholar] [CrossRef] [PubMed]
  12. Muckelt, P.E.; Roos, E.M.; Stokes, M.; McDonough, S.; Grønne, D.T.; Ewings, S.T.; Skou, S. Comorbidities and their link with individual health status: A cross-sectional analysis of 23,892 people with knee and hip osteoarthritis from primary care. J. Comorb. 2020, 10, 2235042x2092045. [Google Scholar] [CrossRef] [PubMed]
  13. Long, H.; Cao, R.; Yin, H.; Yu, F.; Guo, A. Associations between obesity, diabetes mellitus, and cardiovascular disease with progression states of knee osteoarthritis (KOA). Aging Clin. Exp. Res. 2023, 35, 333–340. [Google Scholar] [CrossRef] [PubMed]
  14. Hunter, D.J.; March, L.; Sambrook, P.N. Knee osteoarthritis: The influence of environmental factors. Clin. Exp. Rheumatol. 2002, 20, 93–100. [Google Scholar]
  15. Rafhi, E.; Stupans, I.; Stevens, J.E.; Park, J.S.; Wang, K.N. The influence of beliefs and health literacy on medication-related outcomes in older adults: A cross-sectional study. Res. Soc. Adm. Pharm. 2025, 21, 47–55. [Google Scholar] [CrossRef]
  16. Zhang, W.; Nuki, G.; Moskowitz, R.; Abramson, S.; Altman, R.; Arden, N.; Bierma-Zeinstra, S.; Brandt, K.; Croft, P.; Doherty, M.; et al. OARSI recommendations for the management of hip and knee osteoarthritis. Part III: Changes in evidence following systematic cumulative update of research published through January 2009. Osteoarthr. Cartil. 2010, 18, 476–499. [Google Scholar] [CrossRef]
  17. Razzaque, M.A.; Islam, M.N.; Devnath, N.; Ferdous, N.; Nazrul, F.B.; Rasker, J.J. Osteoarthritis of the knee refractory to standard care: Comparison of intraarticular hylan G-F 20 with triamcinolone acetonide. A real life study in a developing country. Curr. Orthop. Pract. 2018, 29, 329–336. [Google Scholar] [CrossRef]
  18. Kolasinski, S.L.; Neogi, T.; Hochberg, M.C.; Oatis, C.; Guyatt, G.; Block, J.; Callahan, L.; Copenhaver, C.; Dodge, C.; Felson, D.; et al. 2019 American College of Rheumatology/Arthritis Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee. Arthritis Rheumatol. 2020, 72, 220–233. [Google Scholar] [CrossRef]
  19. da Costa, B.R.; Reichenbach, S.; Keller, N.; Nartey, L.; Wandel, S.; Jüni, P.; Trelle, S. Retraction and republication—Effectiveness of non-steroidal anti-inflammatory drugs for the treatment of osteoarthritis pain: A network meta-analysis. Lancet 2017, 390, p109, Retraction in Lancet 2017, 390, e21–e33.. [Google Scholar] [CrossRef]
  20. Stein, C.; Reinecke, H.; Sorgatz, H. Opioid use in chronic noncancer pain: Guidelines revisited. Curr. Opin. Anaesthesiol. 2010, 23, 598–601. [Google Scholar] [CrossRef]
  21. Krebs, E.E.; Gravely, A.; Nugent, S.; Jensen, A.C.; DeRonne, B.; Goldsmith, E.S.; Kroenke, K.; Bair, M.J.; Noorbaloochi, S. Effect of Opioid vs Nonopioid Medications on Pain-Related Function in Patients With Chronic Back Pain or Hip or Knee Osteoarthritis Pain: The SPACE Randomized Clinical Trial. JAMA 2018, 319, 872–882. [Google Scholar] [CrossRef]
  22. Day, A.L.; Curtis, J.R. Opioid use in rheumatoid arthritis: Trends, efficacy, safety, and best practices. Curr. Opin. Rheumatol. 2019, 31, 264–270. [Google Scholar] [CrossRef]
  23. Wang, Y.; Aaron, R.; Attal, N.; Colloca, L. An update on non-pharmacological interventions for pain relief. Cell Rep. Med. 2025, 6, 101940. [Google Scholar] [CrossRef] [PubMed]
  24. Galletta, M.; Piazza, M.F.; Meloni, S.L.; Chessa, E.; Piras, I.; Arnetz, J.E.; D’aloja, E. Patient Involvement in Shared Decision-Making: Do Patients Rate Physicians and Nurses Differently? Int. J. Environ. Res. Public. Health 2022, 19, 14229. [Google Scholar] [CrossRef] [PubMed]
  25. Fransen, M.; McConnell, S.; Harmer, A.R.; Van Der Esch, M.; Simic, M.; Bennell, K.L. Exercise for osteoarthritis of the knee: A Cochrane systematic review. Br. J. Sports Med. 2015, 49, 1554–1557. [Google Scholar] [CrossRef] [PubMed]
  26. Sinatti, P.; Romero, E.A.S.; Martínez-Pozas, O.; Villafañe, J.H. Effects of Patient Education on Pain and Function and Its Impact on Conservative Treatment in Elderly Patients with Pain Related to Hip and Knee Osteoarthritis: A Systematic Review. Int. J. Environ. Res. Public Health 2022, 19, 6194. [Google Scholar] [CrossRef]
  27. Rugelbak, G.M.; Torstensen, T.A.; Paulsberg, F.; Grooten, W.J.A.; Østerås, H. Does exercise adherence influence outcome in knee osteoarthritis?—A secondary analysis of a superiority randomized controlled trial. Physiother. Theory Pract. 2025, 1–5. [Google Scholar] [CrossRef]
  28. Büchel, C. The role of expectations, control and reward in the development of pain persistence based on a unified model. eLife 2023, 12, e81795. [Google Scholar] [CrossRef]
  29. Langford, D.J.; Lou, R.; Sheen, S.; Amtmann, D.; Colloca, L.; Edwards, R.R.; Farrar, J.T.; Katz, N.P.; McDermott, M.P.; Reeve, B.B.; et al. Expectations for Improvement: A Neglected but Potentially Important Covariate or Moderator for Chronic Pain Clinical Trials. J. Pain 2023, 24, 575–581. [Google Scholar] [CrossRef]
  30. Waardenburg, S.; de Meij, N.; van Kuijk, S.M.J.; Van Zundert, J.; Brouwer, B.A. Expectations of treatment outcome in complex cases of patients with chronic pain: A study on the DATAPAIN cohort. Pain Pract. 2024, 24, 8–17. [Google Scholar] [CrossRef]
  31. Degerstedt, Å.; Alinaghizadeh, H.; Thorstensson, C.A.; Olsson, C.B. High self-efficacy—A predictor of reduced pain and higher levels of physical activity among patients with osteoarthritis: An observational study. BMC Musculoskelet. Disord. 2020, 21, 380. [Google Scholar] [CrossRef]
  32. Sale, J.E.M.; Gignac, M.; Hawker, G. The relationship between disease symptoms, life events, coping and treatment, and depression among older adults with osteoarthritis. J. Rheumatol. 2008, 35, 335–342. [Google Scholar] [PubMed]
  33. Wallis, J.A.; Taylor, N.F.; Bunzli, S.; Shields, N. Experience of living with knee osteoarthritis: A systematic review of qualitative studies. BMJ Open 2019, 9, e030060. [Google Scholar] [CrossRef] [PubMed]
  34. Dwek, M.R.; Mcbain, H.; Cleanthous, S.; Shipley, M.; Newman, S. Osteoarthritis: Patient Expectations about Future Pain, Stiffness and Function. Musculoskelet. Care 2015, 13, 84–92. [Google Scholar] [CrossRef] [PubMed]
  35. Martinez-Calderon, J.; Flores-Cortes, M.; Morales-Asencio, J.M.; Luque-Suarez, A. Which Psychological Factors are Involved in the Onset and/or Persistence of Musculoskeletal Pain? An Umbrella Review of Systematic Reviews and Meta-analyses of Prospective Cohort Studies. Clin. J. Pain 2020, 36, 626–637. [Google Scholar] [CrossRef]
  36. Flores-Cortés, M.; Stanton, T.R.; Cuenca-Martínez, F. The Role of Patient Expectations in Treatment Outcome and Satisfaction in Osteoarthritis: A Scoping and Mapping Review. J. Clin. Med. 2025, 14, 8440. [Google Scholar] [CrossRef]
  37. Stanton, T.R.; Braithwaite, F.A.; Butler, D.; Moseley, G.L.; Hill, C.; Milte, R.; Ratcliffe, J.; Maher, C.; Tomkins-Lane, C.; Pulling, B.W.; et al. The EPIPHA-KNEE trial: Explaining Pain to target unhelpful pain beliefs to Increase PHysical Activity in KNEE osteoarthritis—A protocol for a multicentre, randomised controlled trial with clinical- and cost-effectiveness analysis. BMC Musculoskelet. Disord. 2021, 22, 738. [Google Scholar] [CrossRef]
  38. Trafimow, D. Some Implications of Distinguishing Between Unexplained Variance That Is Systematic or Random. Educ. Psychol. Meas. 2017, 78, 482. [Google Scholar] [CrossRef]
  39. Nguefack, H.L.N.; Pagé, M.G.; Katz, J.; Choinière, M.; Vanasse, A.; Dorais, M.; Samb, O.M.; Lacasse, A. Trajectory Modelling Techniques Useful to Epidemiological Research: A Comparative Narrative Review of Approaches. Clin. Epidemiol. 2020, 12, 1205–1222. [Google Scholar] [CrossRef]
  40. Alqahtani, B.A.; Alenazi, A.M. Multisite Musculoskeletal Pain Is Associated with Long-Term Declined Physical Quality of Life and Knee-Related Quality of Life in Older Adults with or at Risk of Knee Osteoarthritis. Medicina 2024, 60, 1305. [Google Scholar] [CrossRef]
  41. Morales-Espinoza, E.M.; Kostov, B.; Salami, D.C.; Perez, Z.H.; Rosalen, A.P.; Molina, J.O.; Paz, L.G.-D.; Momblona, J.M.S.; Àreu, J.B.; Brito-Zerón, P.; et al. Complexity, comorbidity, and health care costs associated with chronic widespread pain in primary care. Pain 2016, 157, 818–826. [Google Scholar] [CrossRef]
  42. Oliveira, C.B.; Maher, C.G.; Franco, M.R.; Kamper, S.J.; Williams, C.M.; Silva, F.G.; Pinto, R.Z. Co-occurrence of Chronic Musculoskeletal Pain and Cardiovascular Diseases: A Systematic Review with Meta-analysis. Pain Med. 2020, 21, 1106–1121. [Google Scholar] [CrossRef] [PubMed]
  43. Kamps, A.; Runhaar, J.; de Ridder, M.; de Wilde, M.; van der Lei, J.; Zhang, W.; Prieto-Alhambra, D.; Englund, M.; de Schepper, E.; Bierma-Zeinstra, S. Occurrence of comorbidity following osteoarthritis diagnosis: A cohort study in the Netherlands. Osteoarthr. Cartil. 2023, 31, 519–528. [Google Scholar] [CrossRef] [PubMed]
  44. Riley, R.D.; Hayden, J.A.; Steyerberg, E.W.; Moons, K.G.M.; Abrams, K.; Kyzas, P.A.; Malats, N.; Briggs, A.; Schroter, S.; Altman, D.G.; et al. Prognosis Research Strategy (PROGRESS) 2: Prognostic factor research. PLoS Med. 2013, 10, e1001380. [Google Scholar] [CrossRef] [PubMed]
  45. Bandura, A. Self-Efficacy: Toward a Unifying Theory of Behavioral Change. Psychol. Rev. 1977, 84, 191–215. [Google Scholar] [CrossRef]
  46. Foster, N.E.; Thomas, E.; Hill, J.C.; Hay, E.M. The relationship between patient and practitioner expectations and preferences and clinical outcomes in a trial of exercise and acupuncture for knee osteoarthritis. Eur. J. Pain 2010, 14, 402–409. [Google Scholar] [CrossRef]
  47. Vase, L.; Wartolowska, K. Pain, placebo, and test of treatment efficacy: A narrative review. Br. J. Anaesth. 2019, 123, e254–e262. [Google Scholar] [CrossRef]
  48. Carrière, J.S.; Pimentel, S.D.; Saba, S.B.; Boehme, B.; Berbiche, D.; Coutu, M.-F.; Durand, M.-J. Recovery expectations can be assessed with single-item measures: Findings of a systematic review and meta-analysis on the role of recovery expectations on return-to-work outcomes after musculoskeletal pain conditions. Pain 2022, 164, e190. [Google Scholar] [CrossRef]
  49. Nicholas, M.K. The pain self-efficacy questionnaire: Taking pain into account. Eur. J. Pain 2007, 11, 153–163. [Google Scholar] [CrossRef]
  50. Roos, E.M.; Klässbo, M.; Lohmander, L.S. WOMAC osteoarthritis index. Reliability, validity, and responsiveness in patients with arthroscopically assessed osteoarthritis. Western Ontario and MacMaster Universities. Scand. J. Rheumatol. 1999, 28, 210–215. [Google Scholar] [CrossRef]
  51. Roos, E.M.; Lohmander, L.S. The Knee injury and Osteoarthritis Outcome Score (KOOS): From joint injury to osteoarthritis. Health Qual. Life Outcomes 2003, 1, 64. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Article metric data becomes available approximately 24 hours after publication online.
Download PDF
Therapeutic Effect of Educational Robotics on Neurocognitive and Social Skills in Two Case Studies of Children with Neurodevelopmental Difficulties
Previous