Quality of life in adolescents with longstanding non-traumatic knee pain An analysis of 316 adolescents with Patellofemoral Pain and Osgood-Schlatter Disease

Objectives: To describe the Quality of Life (QoL) among adolescents with Patellofemoral Pain (PFP) and Osgood-Schlatter Disease (OSD) and investigate characteristics associated with QoL. Study design: Cross-sectional. Participants: 316 adolescents with PFP or OSD. Main outcome measures: QoL subscale of The Knee injury and Osteoarthritis Outcome Score (KOOS) and the EuroQol 5-dimensions (EQ-5D). Results: The KOOS-QoL was 51 ± 18, and the total index score for the EQ5D was 0.67 ± 0.21. KOOS-QoL subscale showed that 60% reported being aware of their knee problems daily or constantly, 38% reported severe to extreme lack of con ﬁ dence in their knees, 28% reported severe to extreme dif ﬁ culty with their knees, and 20% reported severely or totally modifying their lifestyle to avoid potentially damaging activities to their knee. EQ-5D showed that 79% experienced problems with everyday activities, 48% re-ported mobility problems, 17% felt worried, sad, or unhappy, and 7% reported problems looking after themselves. Conclusions: Many adolescents with longstanding non-traumatic knee pain experience low QoL. More than half were aware of their knee problems at least daily, one in three reported a severe lack of con-ﬁ dence in their knee, and one in six felt worried, sad, or unhappy. © 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

A recent individual participants meta-analysis investigating prognostic factors for long-term outcomes demonstrated that 51% of adolescents with non-traumatic knee pain still report knee pain after 12 months (Holden et al., 2021).Notably, low Health-Related Quality of Life (HRQoL) was one of the strongest prognostic factors for poor outcomes and the only identified construct that seems modifiable (Holden et al., 2021).Importantly, modifiable risk factors have high clinical value, as they might help clinicians tailor patient-centered management strategies to enhance rehabilitation success (Murillo et al., 2022).However, it is challenging to identify characteristics of predictive value, and previous evidence regarding prognostic factors in adolescents CMP is sparse and inconsistent (Holden et al., 2021;Huguet et al., 2016;Pate et al., 2020).
Despite the fact that HRQoL has been suggested as an important modifiable prognostic factor in this population, it has remained unclear which characteristics are associated with this phenomenon (Holden et al., 2021;Huguet et al., 2016).However, QoL is a multidimensional construct that is highly subjective and associated with many aspects in adolescents experiencing CMP, such as stress, loneliness, lower self-efficacy, and lower self-esteem (Karimi & Brazier, 2016;Losada-Puente, Araújo, & Muñoz-Cantero, 2020;Mikkelsen et al., 2021).Additionally, identifying ways to manage and improve QoL is ranked as a high research priority among children with CMP, which emphasizes the need for an increased understanding of what influences QoL in adolescents diagnosed with PFP or OSD (Birnie et al., 2019).This will strengthen our knowledge of this phenomenon and plausibly provide important modifiable targets for current management strategies (Pate et al., 2020).
Existing data from three published original trials on adolescents with PFP and OSD foster the opportunity to explore QoL and influencing characteristics in this population (Rathleff et al., 2020b;Rathleff, Graven-Nielsen et al., 2019).Combining data from these three trials increases the sample size, thus strengthens statistical power.Therefore, we aimed to 1) describe QoL among adolescents with PFP or OSD according to the EuroQol 5-dimensions (EQ-5D) and The Knee injury and Osteoarthritis Outcome Score (KOOS) QoL and 2) investigate measures of QoL and its association with characteristics in terms of sports participation, sports activities per week, pain duration, pain intensity, bilateral knee pain, body mass index (BMI), sex, age, and use of pain medication.

Study design
Our study is reported in accordance with the Strengthening the reporting of observational studies in epidemiology (STROBE) statement (Elm et al., 2007).This cross-sectional explorative study was designed as an individual participant analysis of three previous clinical trials on adolescents (defined by the World Health Organization (WHO) as the second decade of life, 10e19 years old (Singh, Siddiqi, Parameshwar, & Chandra-Mouli, 2019)) suffering from longstanding non-traumatic knee pain.Trials are registered on clinicaltrials.gov(NCT02799394, NCT02402673, NCT01438762).Signed informed consent was obtained from all participants, and all data were stored on a secure fileshare.This study was initiated by the original author group of the original publications (Rathleff, Roos, Olesen, & Rasmussen, 2015, 2020b;Rathleff, Graven-Nielsen et al., 2019).

Patient demographics and exposures
All studies collected information on sports participation, sports activities per week, pain duration, bilateral knee pain, BMI, sex, age, use of pain medication, and knee pain intensity (worst pain last week) using a visual analog scale or numeric pain rating scales (measured on a 0e100 scale with 0 being "no pain" and 100 being "the worst pain imaginable").
Knee-related QoL was collected by the KOOS questionnaire, which contains five subscales (pain, symptoms, activity in daily living, function in sport and recreation, and knee-related QoL) scored from 0 to 100 (worst to best) (Collins, Misra, Felson, Crossley, & Roos, 2011;Roos, Roos, Lohmander, Ekdahl, & Beynnon, 1998).As this analysis's primary interest was to describe QoL, we collected data from the KOOS-QoL subscale and individual items representing the knee-related QoL.Health-related QoL was measured by the adult version of EQ-5D in study one (Rathleff et al., 2015) and the EQ-5D youth version in study two and three (Rathleff et al., 2020b;Rathleff, Graven-Nielsen et al., 2019).The two versions contain the same items but slightly different wordings to ensure adolescents below the age of 15 also comprehend the language used.The Danish weights were used to calculate the EQ-5D index score (a higher index score equals better QoL) (EuroQol Research Foundation.EQ-5D-Y User Guide, 2020, n.d.).

Data harmonization and checking
After acquiring the data, we relabeled individual data sets and recoded them to ensure consistency and construct a single, merged dataset.MSR and TY merged the dataset and compared it with the individual dataset and previous publications.The individual patient data validity was checked by undertaking completeness and consistency checks on individual participant data to identify missing or invalid (e.g., out of range) items.Missing information or inconsistencies were checked and rectified as necessary.The original three studies were powered to explore the clinical benefits of a specific management strategy.The sample size was thus fixed by the original studies.

Data analysis
Normally distributed continuous variables are described as mean ± standard deviation.Non-normally distributed data is presented as median, interquartile range, and [range].Categorical variables are expressed as frequencies and percentages.
For Knee-related QoL, we calculated the subscale mean score measured by the KOOS-QoL as recommended, where points are summed, converted to percentages of a maximum score, and reversed to a 0e100, worst to best scale (Collins et al., 2011;Roos et al., 1998).In addition, the individual items were divided into scores from Likert scales (Collins et al., 2011;Roos et al., 1998).All items were answered on a 5-point Likert scale (never/not at all, monthly/mildly, weekly/moderately, daily/severely, constantly/ totally/extreme), yielding 0e4 points, with 0 points representing no difficulty (Collins et al., 2011;Roos et al., 1998).For the QoL measured by the EQ-5D, we present the index score as a general proxy for QoL, the "health profiles", and the individual five items divided into "no problems" or "any problems", as suggested in the EQ-5D User Guide (EuroQol Research Foundation.EQ-5D-Y User Guide, 2020, n.d.).
Characteristics associated with the KOOS-QoL and EQ-5D are presented using box plots, displaying the range, median, and interquartile range of scores.The horizontal line represents the median score, and each box represents the interquartile range.The "whiskers" signify the lowest and highest score for each group.The circles outside the whiskers represent scores above the 10th and 90th percentiles, respectively; these scores fall outside the regular range of values in a distribution and are regarded as outliers.For pragmatic reasons, we divided BMI into five groups (1: <18.5, 2: 18.5 to 24.9, 3: 25 to 29.9, 4: 30 to 39.9, and 5: 40) according to the description provided by the WHO (de Onis et al., 2007).No participants had a BMI >40; therefore, this category is not represented.We divided knee pain duration into three categories (1: 0e6 months, 2: 6e12 months, and 3: 12 months), as recommended (Barton et al., 2021), and lastly, knee pain-intensity was divided into quartiles (Q1: 0-0.25, Q2: 0.25e0.5,Q3: 0.5e0.75, and Q4: 0,75e1).All analyses were carried out using Statistical Package for the Social Sciences (SPSS), version 25 (IBM Corp. Released 2017.IBM SPSS Statistics for Windows, Version 25.0.Armonk, NY: IBM Corp).Based on visual interpretation (median and interquartile range), we examined the association between measures of QoL and included variables (Portney, 2020).As this was an explorative study, we avoided specific hypothesis testing.We interpreted the estimates and figures and looked for differences between subgroups that would be clinically meaningful (Portney, 2020).

Characteristics of participants
A total of 316 participants were included in the dataset; 237 (75%) were girls, 241 (76%) reported bilateral knee pain, and the median age was 14, ranging from 10 to 19 years of age (Table 1).Of all participants, 243 (77%) engaged in sports activities at least twice a week, and 66 (21%) used pain medication due to knee pain.

KOOS-QoL subscale scores
The mean (±SD) KOOS-QoL subscale score for the 316 participants was 51 ± 18.The most prominent item of KOOS-QoL was awareness of the knee problem, with 191 (60%) reporting daily or constant awareness (Table 2).Secondly, 119 (38%) adolescents reported a severe to extreme lack of confidence in their knee, and 87 (28%) reported severe to extreme difficulty with their knees.Lastly, 64 (20%) reported severely or totally modifying their lifestyle to avoid potentially damaging activities.

Characteristics associated with QoL
Visual inspection showed no strong linear association between weekly sports participation and KOOS-QoL or EQ-5D scores (Fig. 1A).Adolescents who participated in little or no sport and those who participated six times weekly or more reported the lowest QoL.For pain duration, no association was observed at any time-points (Fig. 1B).Visual inspection revealed no association between knee pain intensity and KOOS-QoL or EQ-5D scores (Fig. 1C).Adolescents reporting use of pain medication did not score lower QoL compared to adolescents who reported no use of pain medication (Fig. 1D).Lastly, adolescents who participated in sports activities reported similar QoL to those adolescents who weren't engaged in sports activities (Fig. 1E).
Sex and age were not associated with lower QoL (Fig. 2A and B).For BMI, we observed no association in measures of QoL and any of the predefined groups (Fig. 2C).Lastly, adolescents with bilateral knee pain reported similar levels of QoL as those with unilateral knee pain (Fig. 2D).

Discussion
We found that a large proportion of adolescents reported a lack of confidence in their knees, often feeling sad or worried, and modified their usual activities due to knee pain.Our findings contribute to the growing evidence that PFP and OSD are associated with negative psychological impact (Holden et al., 2021;Johansen et al., 2022;Rathleff et al., 2020a).These findings extend our current knowledge and highlight the negative impact longstanding non-traumatic knee pain can have on adolescents, which emphasizes the complex experience of CMP (Khanom et al., 2020;Sørensen & Christiansen, 2017).This may prompt clinicians with new targets during rehabilitation, beyond the functional deficits these young individuals diagnosed with PFP or OSD may experience.

Explanation of findings
Of the 316 adolescents, 74% changed their usual activities to avoid what they perceived as damaging to their knees, and 38% reported a severe to extreme lack of confidence in their knees.Our findings are consistent with a previous study among adolescents diagnosed with either PFP or OSD, which showed that more than 50% reported their reduced sports participation due to "pain", and "I am afraid to damage my knee" (Rathleff et al., 2020a).Adolescents understanding of their pain condition may directly impact how they interpret and manage their pain (Fegran et al., 2021;Khanom et al., 2020;Sørensen & Christiansen, 2017).Some adolescents with a lived experience of CMP ascribe positive changes in their understanding of pain (i.e., pain as non-threatening) as important contributors towards their recovery (Fegran et al., 2021;Joslin, Donovan-Hall, & Roberts, 2021;Khanom et al., 2020).Johansen et al. recently showed that the individual perceptions of 'why do I have pain' (i.e., functional theories) strongly drive behavior in young adults with non-traumatic knee pain, underlining the importance of knowledge about their condition (Johansen et al., 2022).Adolescents with more knowledge about the causes of their pain also showed an increased ability to accept, cope and control their knee pain (Johansen et al., 2022).These psychological processes of change may explain why QoL is associated with prognosis (Holden et al., 2021;Johansen et al., 2022;Khanom et al., 2020).Combined with our findings, it might suggest that psychological processes, such as negative beliefs about pain, play an important role in rehabilitation for adolescents with longstanding non-traumatic knee pain.
Our findings raise important questions about current rehabilitation for adolescents experiencing longstanding non-traumatic knee pain.Traditionally, PFP and OSD management has focused on physical interventions, such as exercise, resistance training, taping, and stretching (Ciatawi & Dusak, 2022;Smith et al., 2017;Willy et al., 2019;Winters et al., 2020).However, it is unclear if these interventions target the QoL dimensions we have identified (Eccleston et al., 2020;Neal, Bartholomew, Barton, Morrissey, & Lack, 2022;Winters et al., 2020).Previous research on  25) 66 (20.9)Sport activities (yes), n (%)  81 (66.9) 38 (74.5) 124 (86.1) 243 (76.9)Sport activities per week, median [range]  2 Continuous variables are presented as median [range].Nominal variables are presented as number (percentages).Abbreviations; NRS, numeric rating scale; VAS, visual analogue scale; (0e100) 0 being "no pain" and 100 being "the worst pain imaginable".adolescents with non-traumatic knee pain show that over the course of treatment, QoL does not improve to the same extent as pain and function (Rathleff et al., 2015(Rathleff et al., , 2018;;Rathleff, Graven-Nielsen et al., 2019).This raises the question of whether we need to reconsider content and focus during rehabilitating adolescents with PFP or OSD (Birnie, Ouellette, Amaral, & Stinson, 2020;Eccleston et al., 2020).Our findings suggest that evidence-based management for adolescents with PFP and OSD may benefit from a broader range of modifiable targets, such as understanding of pain, low confidence, negative emotions, and fear of damaging the knees to improve QoL and potentially their prognosis.Nonetheless, we cannot exclude any possibilities that other relevant contributing dimensions may play a significant role during rehabilitation since these were not measured.The cross-sectional design of this study prevents us from both causal conclusions and a longitudinal understanding; we urge future studies to examine this further.No problems 106 (87.6) 40 (78.4) 115 (79.9) 261 (82.6)Some problems 14 (11.6) 9 (17.6) 26 (18.1) 49 (15.5)A lot of problems 1 (0.8) 2 (3.9) 3 (2.1)6 (1.9) Missing 0 0 7 7 EQ5D Index-score [Mean ±SD]  0.74 ± 0.12 0.63 ± 0.23 0.62 ± 0.25 0.67 ± 0.21 Descriptive data are [n (%)] unless otherwise stated.Values in the bottom row (EQ5D Index-score) are mean ± SD.EQ-5D index score (0e100) where a higher index score equals better QoL.Abbreviations; EQ5D, EuroQol 5-dimensions; SD, Standard deviation.In the current study, we observed that over half (60%) of adolescents with PFP or OSD reported daily or constant awareness of their knee pain, and almost one-fifth felt worried, sad, or unhappy.However, these negative beliefs and important dimensions of QoL were not associated with the duration of knee pain, although this is a well-known determinant for prognosis (Holden et al., 2021).Measures of QoL have been suggested as a proxy for the current state of life in relation to self-evaluated health and, therefore, are prone to a significant change in meaning over time (Schwartz & Rapkin, 2004;Vanier et al., 2021).Adolescents' preferences and goals change over the course of time and are highly dependent of the perceived stage of rehabilitation (Johansen et al., 2022;Joslin et al., 2021;Weiss et al., 2013).Sprangers and Schwartz describe this phenomenon as response shift, a result of changing personal norms, values, or conceptualizing on QoL throughout the course of a condition (Sprangers & Schwartz, 1999;Vanier et al., 2021).The response shift illustrates individual capacity to adapt to their context (i.e., acceptance of pain), shifting focus away from what one cannot (for now) to what one can do to exert control of pain (Sprangers & Schwartz, 1999;Vanier et al., 2021).This aligns with qualitative evidence that suggests some adolescents with CMP ascribe enhanced knowledge of pain as a legitimization of helpful self-management behaviors (e.g., accepting limitations, justifying actions to peers, and prioritizing new life goals) (Fegran et al., 2021;Johansen et al., 2022;Joslin et al., 2021).These processes of change may expedite satisfactory levels of perceived QoL (Johansen et al., 2022;Joslin et al., 2021;Ruskin et al., 2017).The temporal and adaptive nature of perceived QoL may explain why this phenomenon differs from other outcomes in the long-term (Vanier et al., 2021).
In our study, we did not find any strong association between measures of QoL and characteristics.However, a common theme across our results were the large data variability, which emphasizes the heterogeneity of self-reported QoL in this population.This is consistent with previous research on traumatic knee injuries in young individuals (Truong et al., 2020).They showed that QoL was very variable and that several individual psychological, social, and contextual dimensions related to QoL influenced and changed differently across all stages during rehabilitation (Truong et al., 2020).These issues reflect the diverse impact and needs during rehabilitation these young individuals may experience.Adolescents' diverse clinical presentation may provide a plausible explanation for current insufficient evidence of recommending specific treatments in the context of non-traumatic knee pain (Neal et al., 2022;Winters et al., 2020).It has been suggested that current management for adolescents' CMP may benefit from a higher degree of personalization that fit into each individual context (Eccleston et al., 2020;McCracken, Yu, & Vowles, 2022).The current study adds new and important insights of psychological processes of change (e.g., understanding of pain, lack of confidence in the knee) into the field of adolescents suffering from PFP or OSD.Our results may help tackle some of the challenges clinicians face when tailoring rehabilitation to fit an individual adolescent.

Clinical implications
Our findings may help increase clinicians' awareness of psychological processes in adolescents diagnosed with PFP or OSD.It is our clear recommendation that clinicians keep our results in mind when providing personalized evidence-based pain assessment and subsequent management for adolescents with PFP or OSD.There may be unique management options to consider in clinical settings.Future management of care might benefit from clinicians addressing the whole biopsychosocial perspective, tailored to the individual's context.Our results highlighted the presence of unique processes of change, such as understanding of pain, which clinicians may consider addressing in the care they provide (Eccleston et al., 2020;Neal et al., 2022;Winters et al., 2020).Lancet Child & Adolescent Health Commission emphasized the importance of intervening earlier to ensure young people get the proper management at the right time (Eccleston et al., 2020).Early identification of these modifiable targets (e.g., processes of change) may help clinicians tailor personalized rehabilitation to optimize relevant outcomes, such as adolescents' perceived QoL (Eccleston et al., 2020).

Strength and limitations
Strengths of our study include the large sample, the broad age range of participants, and the use of two complementary measures of QoL, one knee-specific and one generic measure, which yielded similar results (Haraldstad et al., 2019).We highlight the fact that psychological processes are highly prevalent in these young individuals and deeply rooted in measures of perceived QoL.Our results may downplay the divisions of therapies (e.g., education vs. exercises) and instead encourage researchers and clinicians to embrace processes of change that are known to drive important outcomes, such as QoL (Holden et al., 2021).The following limitations should also be considered in interpreting the findings of our study.First, our results rely on a cross-sectional explorative study design.Therefore, our results should be seen as hypotheses generating rather than proving any causal or temporal relationships between measures of QoL and characteristics.Second, recent studies emphasize that current measures of QoL are prone to conceptual and methodological challenges and may lack the validity of the target construct (Hansen et al., 2022;Haraldstad et al., 2019;Le et al., 2021).Therefore, our data must be interpreted cautiously as we may have missed important dimensions of QoL (Haraldstad et al., 2019;Krogsgaard & Hansen, 2022;Le et al., 2021;Vanier et al., 2021).Third, we used visual interpretation to analyze the data exclusively.In most situations, a combination of visual and statistical analyses may provide the clearest interpretation (Portney, 2020).

Future research
Using qualitative methods, future studies should explore the questions of QoL to capture the entire experience of longstanding non-traumatic knee pain and understand how and why changes in perception occur over time.We encourage future studies to use a longitudinal design to understand how QoL can be modified and explore co-occurring dimensions associated with QoL as well as changes in QoL.This will provide essential knowledge to further understand the complexity of living well with (or despite) knee pain and potentially develop new management strategies for these individuals.

Conclusion
A high proportion of adolescents with longstanding nontraumatic knee pain experience low QoL.More than half were aware of their knee problems at least daily, one in three reported a severe lack of confidence in their knee, and one in six felt worried, sad, or unhappy.Our results may encourage clinicians to extend the scope and address psychological processes of change in rehabilitation for adolescents with longstanding non-traumatic knee pain.

Name of the public trials registry and the registration number
ClinicalTrials.govIdentifier: NCT02799394, NCT02402673, NCT01438762.

Fig. 1 .
Fig. 1.Association between measures of QoL (EQ-5D index score and KOOS QoL scores) and characteristics.(A) Sports activities per week; (B) Knee pain duration; (C) Knee pain intensity; (D) Pain medication; (E) Sports participation.The horizontal line represents the median score, and each box represents the interquartile range.The "whiskers" signify the lowest and highest score for each group.The circles outside the whiskers represent scores above the 10th and 90th percentiles.The error bars represent the.Abbreviations; EQ5D, EuroQol 5-dimensions; KOOS, Knee injury and Osteoarthritis Outcome Score; Q, Quartiles; QoL, Quality of life.

Fig. 2 .
Fig. 2. Association between measures of QoL (EQ-5D index score and KOOS QoL scores).(A) Sex; (B) Age; (C) Body mass index (BMI); (D) Bilateral knee pain.The horizontal line represents the median score, and each box represents the interquartile range.The "whiskers" signify the lowest and highest score for each group.The circles outside the whiskers represent scores above the 10th and 90th percentiles.Abbreviations; EQ5D, EuroQol 5-dimensions; KOOS, Knee injury and Osteoarthritis Outcome Score; QoL, Quality of life.

Table 1
Characteristics of included participants and studies.

Table 2
Participants Knee injury and Osteoarthritis Outcome Score (KOOS) QoL subscale.Values in the bottom row (KOOS QoL subscale) are mean ± SD.KOOS QoL subscale scored from 0 to 100 (worst to best).Abbreviations; KOOS, Knee injury and Osteoarthritis Outcome Score; SD, Standard deviation; QoL, Quality of life.

Table 3
EQ-5D data, showing frequencies and proportions by dimension and severity level.