Factors associated with restricted mobility outside the home in community-dwelling adults ages fifty years and older with knee pain: An example of use of the International Classification of Functioning to investigate participation restriction
To investigate the links between knee pain characteristics and restricted mobility outside the home, and how these are influenced by mobility-specific activity limitation, age, sex, socioeconomic status, environmental factors, and comorbidity.
We conducted a cross-sectional survey of community-dwelling adults age ≥50 years. A total of 2,252 responders reporting pain in and around the knee in the last year were eligible. The primary outcome was self-reported restricted mobility outside the home in the previous 4 weeks (dichotomized as present or absent).
Knee pain severity was strongly associated with restricted mobility outside the home, an association largely mediated by perceived limitation in walking. After adjusting for demographic and socioeconomic characteristics, individual contributions from selected comorbidities, knee pain severity, limitation in walking, and specific environmental factors remained. These environmental factors included perceived need of aids and assistance (adjusted odds ratio [OR] 3.1, 95% confidence interval [95% CI] 2.2–4.4), poor access to public transportation (adjusted OR 2.3, 95% CI 1.4–3.9), and having no access to a car (adjusted OR 1.6, 95% CI 1.1–2.4).
There are a range of potential health and social targets that, if addressed, might reduce restricted mobility outside the home in middle-aged and older individuals with knee pain. Our results suggest that, in addition to treating the knee symptoms, such targets might include comorbidity, walking ability, and environmental barriers such as poor access to public transportation. Moreover, removing environmental barriers may reduce immobility outside the home even in the continued presence of osteoarthritis symptoms and specific activity limitations.
Osteoarthritis is the most common cause of joint pain and disability in older persons (1), but the ways in which it leads to disability are complex. In the World Health Organization's International Classification of Functioning, Disability and Health (ICF), disability is an umbrella term for impairments (e.g., joint pain, stiffness), activity limitation (e.g., difficulties walking), and participation restriction (e.g., restricted mobility outside the home), which refer to the consequences of health conditions at tissue or organ, individual, and societal levels, respectively (2). Evaluation of all 3 levels and their interactions will help to characterize the impact of osteoarthritis (3).
Population studies of the impact of knee osteoarthritis, for example, have focused on activity limitation (4–8). The extent to which these studies also measure participation restriction is unclear. Participation restriction is defined as “problems an individual may experience in involvement in life situations” (2). A common sense view is that knee pain might cause specific limitation in walking (an activity limitation), and that such limitation directly restricts a person's capacity to get around outside the home. The assumption of a direct path from impairment to participation restriction via specific activity limitation was criticized in the earlier version of the ICF for implying that only medical approaches are necessary to manage health (9, 10). However, it may be perfectly feasible to maintain participation despite continuing impairment (such as knee pain) and limitation of a specific activity (such as walking) (11); for example, availability of transportation or use of a walking frame, or contentment that others do the shopping, may mean that getting around outside the home is satisfactory despite persistent symptoms.
The extent to which older persons with joint problems perceive themselves as being restricted has been less extensively investigated than joint pain severity and associated specific functional limitations. There is evidence of a link between valued life roles and psychological well-being (12); social context shapes the impact of a health condition on a person's life (13) and may be of more concern to individuals than impairments and activity limitations (14), and there is a growing acknowledgment that environmental factors may explain the gap between an individual's capacity and his or her actual performance in activities of daily life (15–17).
We have previously found that adults with joint pain age 50 years and older in the general population are more likely to report participation restriction than those without joint pain, and that the most common form of participation restriction in this age group is restricted mobility outside the home. In this new study, we investigated the extent to which knee impairments (pain severity, stiffness severity, and chronicity of pain) in older individuals are associated with this specific area of participation restriction (mobility outside the home), and the extent to which such associations are explained by a specific functional limitation (walking). We also investigated 3 specific hypotheses: that 1) environmental factors, such as public and private transportation, and 2) the presence of other health conditions and impairments may be separately associated with this restriction, and 3) environmental factors may further explain the link between limitations in walking and restricted mobility outside the home.
PARTICIPANTS AND METHODS
Study design and participants.
Data for this study were collected in the 2-stage baseline survey of the North Staffordshire Osteoarthritis Project, a population-based study of adults ages 50 years and older registered with 3 general practices in North Staffordshire (irrespective of actual consultation) (n = 11,309) (18). Seventy-nine registered adults were excluded prior to mailing (55 due to death or departure from the practice and 24 due to known cognitive difficulties or terminal illness). The remaining 11,230 participants were mailed a Health Survey questionnaire, of which a further 175 were excluded (45 due to death or departure, 105 due to a wrong address, and 25 due to terminal illness) and 7,878 responded; response rate after excluding the ineligible participants after mailing was 71.3%. Responders who gave their consent for further contact and indicated that they had knee pain in the last year (n = 2,863) were sent a Regional Pains Survey questionnaire. A further 4 participants were excluded (3 deaths and departures and 1 due to wrong address) and 2,540 responded to this second questionnaire; response rate after excluding ineligible participants after mailing was 88.8%. Analysis was performed with responders who indicated in both questionnaires that they had knee pain in the last year (n = 2,252). Ethical approval for the study was obtained from the North Staffordshire Local Research Ethics Committee.
The dependent variable was restricted mobility outside the home, measured by one item from the Keele Assessment of Participation (KAP) in the Health Survey questionnaire (19). The KAP measures participation restriction from the perspective of the individual. Items are phrased to capture performance (“I have”) and individual judgment and the nature and timeliness of participation (“as and when I have wanted”). The reliability and validity of the KAP are adequate for providing estimates of perceived participation restriction in population studies (19). Responses are on a 5-point scale (all, most, some, a little, none of the time) and responders were considered to be restricted in their mobility outside the home if they did not get around outside their home during the previous 4 weeks “as and when they wanted” for “all” or “most of the time.”
The independent variables in the analysis represented knee-specific characteristics (collected in the Regional Pains Survey questionnaire), mobility-specific activity limitation, environmental factors, other comorbidities, and socioeconomic factors (all collected in the Health Survey questionnaire). Knee pain and stiffness severity were measured using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (20). The Likert version was used, which offers a 5-point ordinal scale from mild (score of 0) to extreme (score of 4) and provides scores for pain (range 0–20) and stiffness (0–8), with higher scores indicating worse health. For analysis, the knee pain and stiffness severity scores were categorized using quartiles. A single item was used to measure the duration of pain (<3 months or ≥3 months).
Mobility-specific activity limitation was measured by one item from the physical functioning scale of the Medical Outcomes Study Short Form 36: limitation walking half a mile (a lot, a little, not limited at all) (21). Data were collected by single items for 3 environmental factors relevant to mobility outside the home: 1 to measure if responders required assistance or aids (i.e., “During the past 4 weeks have you required the assistance of others or aids to move around outside your home?”) and 2 to measure access to transportation (i.e., “Do you have access to a car when you personally need it?”; “Do you have access to public transport?”). These items had a simple yes/no response option.
The comorbidities included in this analysis were previously found to be independently associated with participation restriction in at least one aspect. These comorbidities were number of peripheral joint pains, number of health conditions, anxiety, depression, and cognitive impairment. The original variable number of peripheral joint pains was replaced by 2 variables to measure the extent of lower limb joint pain derived from separate measures of pain in the hips, knees, and feet in the last 12 months: bilaterality of knee pain and number of other lower limb joint pains (0–4 using the hip and foot questions). The number of health conditions was a simple count of the presence of 3 self-reported health conditions (chest problems, heart problems, and diabetes). Anxiety and depression during the previous week were measured using the Hospital Anxiety and Depression Scale; raw scores were calculated and used to categorize individuals as noncases (score range 0–7), possible cases (score range 8–10), and probable cases (score range 11–21) (22). Cognitive impairment was measured using the cognitive and alertness behavior subscale of the Functional Limitations Profile. Raw scores were categorized according to the distribution of the sample using quartiles; 53% of responders had a score of 0 (forming the “no cognitive impairment” category), and the third quartile (75th percentile: 22.4) separated the rest into high and low levels (23). Body mass index (BMI), calculated from self-reported height and weight, was included in this analysis because of its link with knee osteoarthritis (24). Responders were categorized into standard BMI groups: normal weight (BMI 20–24.9 kg/m2), underweight (BMI <20 kg/m2), overweight (BMI 25–29.9 kg/m2), and obese (BMI ≥30 kg/m2).
The socioeconomic characteristics included were those previously found to be associated with participation restriction in the general population: occupational class (higher professional/managerial, lower professional/managerial, intermediate occupation, self-employed, lower supervisory, semi-routine, routine) (25, 26), educational attainment (those who finished their education on leaving school, those who went on to further education such as college or university), and perceived adequacy of income (quite comfortable, able to manage without much difficulty, have to be careful with money, find it a strain to get by from week to week) (27).
The analysis of associations was performed in 5 stages.
Associations between knee-specific characteristics (pain severity, stiffness severity, and pain duration) and restricted mobility outside the home, unadjusted and adjusted for potential covariates (first for age [in 10-year bands] and sex, and then for socioeconomic status [occupational class and educational attainment]), were examined using logistic regression.
Associations between knee pain severity and restricted mobility outside the home adjusting for limitation walking half a mile and for the potential covariates age, sex, and socioeconomic status were examined using logistic regression.
Associations between restricted mobility outside the home and limitation walking half a mile were examined using cross-tabulation for the entire knee pain sample, and separately for 1) those with versus those without access to public transportation, 2) those with versus those without access to a car, and 3) those requiring versus those not requiring aids/assistance to get around outside the home.
Multivariate analysis of associations between restricted mobility outside the home and knee-specific characteristics, limitation walking half a mile, comorbidities (number of other lower limb joint pains, number of health conditions, bilaterality of knee pain, anxiety, depression, cognitive impairment, BMI), age, sex, socioeconomic status, and environmental factors was conducted using logistic regression.
Interaction terms were added to the multivariate model in stage 4. First, a term was fitted for limitation walking half a mile and each environmental factor (requirement of aids/assistance for mobility outside the home, access to a car, and access to public transportation). Second, a term was fitted for the requirement of aids and assistance and access to public transportation.
Characteristics of the study population.
Of the 2,252 individuals with knee pain in the study population, 1,305 (58%) were female. Their overall mean ± SD age was 65.3 ± 9.7 years. The mean ± SD WOMAC scores were 6.4 ± 4.6 for pain, 2.7 ± 2.0 for stiffness, and 21.4 ± 16.2 for physical functioning. Mean scores for all of these subscales increased with increasing age, and were higher in women. Most participants reported bilateral knee pain (56.3%) and pain in at least one other lower limb joint (70.4%). A total of 712 (32.0%) reported restricted mobility outside the home.
Each knee characteristic was significantly associated with restricted mobility outside the home after adjusting for age group, sex, and socioeconomic status (Table 1). There was a trend of increasing mobility restriction outside the home with increasing levels of pain and stiffness. Adjusting for age, sex, and socioeconomic status, the strongest association was with the highest category of pain severity (adjusted odds ratio [OR] 19.8, 95% confidence interval [95% CI] 13.6–28.9).
Table 1. Associations between restricted mobility outside the home and knee-specific variables*
OR = odds ratio; 95% CI = 95% confidence interval.
Age- and sex-adjusted.
Adjusted for age, sex, and socioeconomic status (occupational class and educational attainment).
Adjusted for age, sex, socioeconomic status, and limitation walking half a mile.
3rd highest (2)
2nd highest (3–4)
3rd highest (3–6)
2nd highest (6.25–10)
Duration of pain, months
Adjustment of the logistic regression model for limitation in walking half a mile substantially reduced the strength of the associations between restricted mobility outside the home and knee pain severity and knee stiffness severity (Table 1), indicating that the associations were largely explained by the specific limitation in walking half a mile (e.g., the OR [95% CI] for the association between the highest category of pain severity and restricted mobility, when adjusted for this limitation, decreased from 19.8 [13.6–28.9] to 4.2 [2.7–6.5]) (Figure 1). Adjustment also caused attenuation in the point estimate for the association with duration of pain (OR 2.8, 95% CI 2.3–3.6 compared with OR 1.7, 95% CI 1.3–2.2).
There was a step up between being limited in walking “a little” and “a lot” in terms of the impact of each of the environmental factors on the strength of associations (Figure 2). Having access to public transportation and access to a car were associated with reduced levels of mobility restriction for all levels of difficulty walking half a mile, including having no difficulty. A total of 165 (35.6%) participants who indicated a lot of difficulty walking half a mile but had access to public transportation did not perceive themselves to be restricted in mobility outside the home, compared with 26 (11.8%) participants who had a lot of difficulty walking half a mile and had no access to public transportation (Table 2). Self-reported requirement of aids or assistance for mobility outside the home was associated with higher levels of mobility restriction outside the home within each level of difficulty walking, including no difficulty.
Table 2. Associations between restricted mobility outside the home and limitation walking half a mile for the knee pain group*
Knee pain sample with restricted mobility outside the home
Difficulty walking half a mile
Limited a little
Limited a lot
Values are the number (percentage restricted).
Access to public transportation
Access to a car
In the multivariate analysis, knee pain severity remained associated with restricted mobility outside the home after adjusting for other health conditions, impairments, limitation walking half a mile, socioeconomic status, and environmental factors (highest category of pain severity: OR 2.4, 95% CI 1.4–4.0) (Table 3). Age (70–79 years: OR 1.6, 95% CI 1.1–2.5, and ≥80 years: OR 2.4, 95% CI 1.3–4.3) was also independently associated with restricted mobility outside the home, but sex and socioeconomic factors were not. All 3 environmental factors (requirement of aids/assistance [OR 3.1, 95% CI 2.2–4.4], poor access to public transportation [OR 2.3, 95% CI 1.4–3.9], and poor access to a car [OR 1.6, 95% CI 1.1–2.4]) remained independently associated with restricted mobility outside the home, as did number of health conditions (2–3 health conditions: OR 2.0, 95% CI 1.2–3.2), depression (possible case: OR 2.3, 95% CI 1.5–3.5, and probable case: OR 3.0, 95% CI 1.7–5.3), cognitive impairment (low level: OR 1.6, 95% CI 1.1–2.4, and high level: OR 1.5, 95% CI 1.0–2.2), and being underweight (OR 3.0, 95% CI 1.5–8.4), but bilaterality of knee pain, number of other lower limb joint pains, and anxiety did not.
Table 3. Associations between restricted mobility outside the home and knee pain, age, sex, socioeconomic status, other comorbidities, limitation walking, and environmental factors*
The interactions between 1) walking limitation and access to a car, 2) walking limitation and access to public transportation, and 3) the requirement for aids and assistance and access to public transportation were not statistically associated with restricted mobility outside the home (data not shown). There was a statistically significant interaction between limitation in walking and reporting the requirement of aids and assistance for mobility outside the home (Table 4). Individuals who required aids or assistance were more likely, at all levels of walking limitation, to be restricted in getting around outside of their home. However, the effect of requiring aids or assistance was less strong in persons who had a lot of difficulty walking. ORs for all other variables in the model remained similar to those in the model without the interaction.
Table 4. Associations between restricted mobility outside the home and interaction between limitation walking half a mile and the requirement for aids and assistance*
OR = odds ratio; 95% CI = 95% confidence interval.
Adjusted for age, sex, knee pain severity, occupational class, educational attainment, adequacy of income, bilateral/unilateral knee pain, number of other lower limb joint pains, number of health conditions, anxiety, depression, cognitive impairment, body mass index, access to public transportation, and access to a car.
Not limited and does not require aids
Limited a little and does not require aids
Limited a lot and does not require aids
Not limited and does require aids or assistance
Limited a little and does require aids or assistance
Limited a lot and does require aids or assistance
In this study, we used the recently proposed ICF framework to structure an investigation of restricted mobility outside the home in community-dwelling adults ages 50 years and older with knee pain. Our findings suggest that restricted mobility outside the home is related to a range of clinical and nonclinical factors in older adults with knee pain.
The severity of knee pain was associated with an increased likelihood of reporting restricted mobility outside the home. This is consistent with the findings of other studies (5, 7, 28), as are the separate but similar associations with mobility restriction that were found for both knee stiffness severity and duration of pain.
Not unexpectedly, we confirmed the strong association between community-dwelling older persons' perceptions that they are restricted in their mobility outside the home if they have knee pain, and their report of a specific limitation in walking. However, we also found a separate link between knee pain and mobility outside the home, which is independent of the capacity to walk. In addition, our findings highlight the important additional effect of environmental factors on the link between limitation walking half a mile and restricted mobility outside the home. Our findings suggest that a simple causal pathway from knee pain via limited walking to a person's sense of restricted mobility does not fully explain the extent of perceived restriction. Environmental factors are separately linked to restricted mobility outside the home in persons with knee pain. For example, having access to public transportation was associated with higher participation levels, suggesting that mobility outside the home is maintained even when knee pain and limitation in walking are present. This provides evidence for the potential benefits of improving the environment in which people live (29), specifically that environmental factors may reduce the disability of osteoarthritis and joint pain (30).
The step up in the link between environmental factors and individuals reporting a little and a lot of limitation walking half a mile suggests that individuals with a lot of limitation walking half a mile are likely to experience the greatest impact of environmental factors. The analysis of interactions between limitation in walking and the requirement for aids/assistance highlights the importance of such aids and assistance. One possible explanation for this interaction is that, in general, individuals are more restricted because they need aids or assistance (which they do not have or are unable to acquire outside the home as and when they want, e.g., they can go out only when it is convenient for the person they rely on for assistance). It is also possible that using aids or assistance is part of how individuals define being restricted (e.g., using a stick is not how I want to get around outside the home). However, individuals who are limited a lot in walking half a mile may be more likely to actually have the aid or assistance that they need. These findings are consistent with the view that requiring aids/assistance is associated with not getting around outside the home as and when individuals want (potential unmet need), and having those aids and assistance facilitates getting around outside as and when they want, especially in persons who have significant difficulty walking (benefits of met need). However, this is a speculative interpretation of questionnaire data.
Comorbidity is also important to restricted mobility outside the home. Depression, for example, was independently associated with restricted mobility and may be a separate target for interventions, as may other comorbidities. Patients with osteoarthritis have higher depression scores than the general population (31) and management of depression can extend beyond reducing depressive symptoms and reduce pain and disability (32). In contrast to other studies of knee pain (33) and disability (34), obesity and being overweight were not as important as being underweight, which may be a marker of frailty (35).
The analyses started with variables known to be associated with participation restriction. This may have excluded some factors pertinent to restricted mobility outside the home in older adults, such as social networks (36) and other health conditions (e.g., vision problems, stroke ), but our method provides a picture of general factors important to perceived mobility in older adults with knee pain. Lack of adjustment for other potential covariates, as well as the interaction between walking limitation and the requirement for aids/assistance, may explain why the impact of aids/assistance was associated with restricted mobility in the multivariate analysis (stage 4).
Our results suggest that environmental factors and other comorbid conditions may explain some of the link between restricted mobility outside the home and limitation walking half a mile in older adults with knee pain. In addition, our study highlights that the oldest groups and the most socioeconomically disadvantaged groups were more likely to have restricted mobility outside the home even after adjusting for comorbidity and environmental factors. In other studies socioeconomic status has been associated with locomotor disability (37) and has been reported to be a predictor of disability in groups of individuals with arthritis (34, 38, 39).
The item used to indicate knee pain covered all grades of knee pain. The profile of the knee pain sample was similar to an earlier study in the same location (8). Some selective nonresponse and nonconsent occurred in our method, which may favor younger and healthier responders. Nonresponse to the Health Survey was greatest in persons ages 80 and older. Nonconsenters with knee pain, compared with those who did participate and were included in this analysis, were older (mean age 66.7 years versus 65.3 years), had worse general health (mean Short Form 12 [SF-12] physical functioning score 36.9 versus 37.4; mean SF-12 mental functioning score 47.2 versus 48.2), and had more restricted mobility (35.9% versus 32.0%). This may have underestimated the strength of associations with increasing age and other comorbidities.
The results of this study are based on the subjective reporting of all variables, other than age and sex. If not participating “as and when I want” relates to actual health care need, the associations observed suggest a number of potential strategies to prevent or restore mobility outside the home. However, this study does not provide evidence of the direction of the cause and effect relationship or the effectiveness of any possible intervention. Both clinical and nonclinical approaches are likely to be needed for maximum effect on participation. Clinical approaches aimed at the symptoms of the knee problem (reducing pain, reducing mobility limitation [either through improving an individual's walking ability or providing a walking aid], and managing other comorbid health conditions such as depression) combined with a social approach aimed at, for example, providing public transportation, or interventions targeted at persons ages 80 and older, may reduce restricted mobility outside the home in older adults with knee pain.
However, we need evidence about actual effectiveness before clear policies for intervention can be established, or optimal targets determined. Our work does present hypotheses to be tested. For example, a reduction in knee pain severity should improve restricted mobility outside the home. It is also possible that reverse causality may occur and a reduction of restricted mobility outside the home may also reduce impairments, such as depression, or encourage people to get around more and prevent further limitation in walking (40). Further prospective studies are required to investigate such ideas and the effectiveness of interventions to reduce this common form of disability in older adults with knee pain.
Dr. Wilkie had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Wilkie, Peat, Thomas, Croft.
Acquisition of data. Wilkie, Peat, Thomas, Croft.
Analysis and interpretation of data. Wilkie, Peat, Thomas, Croft.
We would like to thank the administrative and health informatics staff at Keele University's Primary Care Sciences Research Centre and the doctors and staff of the participating general practices. We would also like to thank Sara Mottram for her comments in drafting this article.