• Self-rated health;
  • Pain;
  • Activity limitations


  1. Top of page
  2. Abstract
  7. Acknowledgements


To longitudinally explore the hypothesized role of worsening pain and development of activity limitations as mediators in the relationship between arthritis onset and worsening self-rated health (SRH).


Data was obtained from the 1998/1999 and 2000/2001 cycles of the population-based Canadian longitudinal National Population Health Survey (n = 10,859; ages ≥18; response rate: time 1 = 81.6%, time 2 = 89.2%). Respondents were asked about chronic conditions, pain, activity limitations, and self-perceived health; change over time was assessed. Change in effect of arthritis onset on worsening SRH upon considering potential mediators was assessed through multivariate logistic regression, controlling for sociodemographic characteristics and onset of other conditions.


Worsening pain fully explained the effect of arthritis onset on worsening SRH; a portion of the effect of worsening pain was mediated by the development of activity limitation. Residual direct effect of arthritis onset was statistically insignificant. Worsening pain and development of activity limitations also mediated a portion of the effects of onset of other chronic conditions but to a lesser extent than arthritis onset.


This is the first study to examine these relationships longitudinally. Identifying the role of mediators is necessary if target areas of prevention and/or management are sought, either at the individual or population level. Our results indicate that the development of arthritis has a significant impact on worsening SRH. Pain and development of activity limitations fully account for the relationship between arthritis onset and worsening SRH. High priority should be placed on prevention and management strategies for pain among people with arthritis.


  1. Top of page
  2. Abstract
  7. Acknowledgements

The use of an overall health status measure, particularly self-rated health (SRH), in population and clinical health surveys has become common (1–4). Historically, much of the work on SRH has focused on its predictive ability for a range of health outcomes, including social-psychological wellbeing (5–7), morbidity (8–10), health care utilization (10–12), and mortality (10, 13, 14). More recently, interest in understanding the determinants of SRH has increased. Predictors of good SRH include younger age (15–17), male sex (15), higher socioeconomic status (15, 16, 18), social support (19, 20), and absence of chronic conditions and pain (15–17, 21–23). Although physiologic and psychological determinants of good SRH have been documented, the main determinant appears to be physical health (15, 24, 25).

The most significant predictor of poor SRH is the presence of chronic conditions (15, 17, 21, 22, 26). However, population-based and clinical studies indicate that 30–80% of persons with at least 1 chronic condition report good health (15–17). This raises questions about the distinction between individuals' perceptions of their health status and a diagnosed medical condition. It has been suggested that some individuals think about specific health problems when asked to rate their health, whereas others think in terms of general physical functioning or health behaviors (27). Pain is a common reason for seeking medical care around the world (28–30), and is associated with poor SRH (15–17, 23). Studies indicate that chronic pain seems to relate to poor health more strongly than chronic diseases or age, and the significant association persists even when sociodemographic and relevant health factors are considered (16, 17).One possible explanation is that pain is a mediator between chronic conditions and SRH. This study explores this idea by examining the onset of arthritis.

All forms of arthritis share the common symptom of pain and have a major impact on functioning and health care utilization (31–38). Concern about arthritis as a public health problem is growing in North America (39), with large projected increases in prevalence as the population ages (37, 40). The population impact of arthritis has been well documented in the literature. The most frequent symptoms reported by adult men and women are pains in the musculoskeletal system (23), and arthritis is the most common cause of disabilities and activity limitations (35, 41). Having arthritis has consistently been shown to be associated with poor SRH (4, 22, 23, 37, 42, 43). Functional limitations associated with musculoskeletal conditions have also been associated with poorer health status (21–23, 42, 43). Previous cross-sectional studies examining the independent effects of several conditions have shown arthritis, and musculoskeletal symptoms generally, to be among the conditions with the strongest associations with reporting poor SRH (23, 42).

Studies examining the determinants of SRH have been predominantly cross-sectional. Few studies have examined predictors of change in SRH, and even fewer have examined the relevance of changes in factors associated with SRH to predicting changes in SRH. This study explores the potential role of worsening pain and development of activity limitation as hypothesized mediators in the relationship between the onset of arthritis and worsening SRH over time by examining data collected from a Canadian cohort over a 2-year period. Given the impact of chronic conditions on SRH generally, the onset of other chronic conditions are considered.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Study design and setting

Data were obtained from the household sample of the 1998/1999 and 2000/2001 cycles of the longitudinal Canadian National Population Health Survey (NPHS) (44). A stratified, multistage design was used for the NPHS, which includes household residents, with some exceptions (44), in all Canadian provinces. The survey collected sociodemographic and in-depth health information from the longitudinal respondent at each cycle. Only individuals ages ≥18 years were asked about the presence of some chronic conditions, therefore only individuals ≥18 years in 1998 were considered eligible for inclusion in this study (n = 12,168). However, due to either incomplete responses, loss to followup, or death, 10,859 individuals remained eligible in 2000/2001. These individuals comprised our study population.

Data were weighted to reflect the sample design, adjustments for nonresponse, and poststratification and are representative of the Canadian household population ages ≥14 years in 1994/1995. The bootstrap method was used for calculating variances of estimates (44).

Variable definitions.

Self-rated health.

Respondents were asked, “In general, would you say your health is: Excellent? Very good? Good? Fair? Poor?” A dichotomous variable was created to indicate worse SRH at followup. Worse was defined as a change in the following direction: excellent/very good to good to fair/poor.

Chronic conditions

In each cycle, respondents were asked, “Do you have any of the following long-term conditions diagnosed by a health professional?” Long-term was defined as having lasted or expected to last ≥6 months, and a list of 19 conditions was presented. A dichotomous variable was created to indicate the onset of arthritis or rheumatism (i.e., negative response in 1998/1999, positive in 2000/2001).

Taking into consideration statistical constraints and distinctions made in the literature (22), the conditions indicated by respondents were placed into 1 of 3 categories: nondisabling (allergies, asthma, migraine headaches, high-blood pressure, epilepsy, stomach or intestinal ulcer, glaucoma, cataracts, urinary incontinence, and thyroid condition), disabling (back problems excluding arthritis, chronic bronchitis or emphysema, Alzheimer's disease or other dementia, and bowel disorder/Crohn's Disease or colitis), and life-threatening (heart disease, cancer, experiencing the effects of a stroke, and diabetes). A count variable for each category was created to indicate the number of conditions developed over the study period.


For the pain and discomfort section of the survey, respondents were told, “The next set of questions asks about your day-to-day health. The questions are not about illnesses such as colds that affect people for short periods of time.” The question asked, “How would you describe the usual intensity of your pain or discomfort?” Respondents were asked to select one of the following categories: none, mild, moderate, or severe. Worse pain was defined as any change towards severe pain/discomfort.

Activity limitations

Respondents were asked if a physical condition or health problem limited the type or amount of activity they could do at home, school, work, or in other activities such as leisure time activities, or whether they had any long-term disabilities or handicaps. A positive response to any one of the above was defined as having an activity limitation. Development of an activity limitation was defined as a change from a negative to a positive response over time.


For descriptive analyses, age was categorized as 18–34 years or subsequent 10-year increments, up to 75+ years. For regression analyses, age was considered continuously.


Based on total household income and the number of individuals living in the household, 4 income categories were derived as recommended by Statistics Canada (44): lowest, lower-middle, upper-middle, and highest. Six percent of respondents did not provide sufficient information and were therefore included in a “not stated” income category.


The respondent's highest level of education, as of 1998, was classified as one of the following: less than secondary school, secondary school graduation, some post-secondary, and post-secondary graduation.

Statistical analyses.

Descriptive analyses.

Proportions of the population with selected characteristics at baseline (SRH status, level of pain, activity limitations, with arthritis, and at least 1 condition for each of the condition categories) were calculated. At followup, the proportion of respondents with worse SRH, worse pain, and those who developed an activity limitation was determined, as well as the proportion with an onset of arthritis and at least 1 condition within each category. Also, the proportion of respondents with each of the followup outcomes who also reported any other followup outcome was documented.

Regression analyses

The longitudinal design of the survey permitted a prospective assessment. However, as measurements were only taken at discrete time points (2-year interval), the actual time of occurrence of events was not available.

To test our conceptual model (Figure 1A), sequential multivariate logistic regression models were estimated. All analyses controlled for age, sex, income, education, and onset of chronic conditions. Three models, deemed “main effects models,” were initially estimated to determine if a relationship existed between arthritis onset and 1) worsening pain, 2) reporting a new activity limitation, or 3) worsening SRH. Three models, deemed “mediated models,” were estimated to establish whether the relationship between onset of arthritis and 1) worsening SRH was mediated by worsening pain, 2) reporting an activity limitation was mediated by worsening pain, and 3) worsening SRH was mediated by development of an activity limitation, adjusted for worsening pain. Mediation was assessed by using a causal steps test of the intervening variable effect described by Baron and Kenny (45). The extent of mediation was quantified as:

thumbnail image

Figure 1. Hypothesized conceptual model (A), main effect models (B), and mediated models (C). All estimates are adjusted for age, sex, education, income, and all new conditions. SRH = self-reported health; OR = odds ratio; NS = not significant. ∗ = main effect; † = adjusted for developing an activity limitation.

Download figure to PowerPoint

the percent change in estimated effects = 100[(ORM− − ORM+)/(ORM− − 1)]

where ORM− = odds ratio for the predictor variable of interest from a regression model unadjusted for the hypothesized mediator and ORM+ = odds ratio for the same predictor variable from a regression model that adjusts for the hypothesized mediator (46).


  1. Top of page
  2. Abstract
  7. Acknowledgements

The prevalence of characteristics of interest at baseline are presented in Table 1. Approximately 10% of respondents reported their health as being fair/poor, ∼15% reported some level of pain, and 17% reported the presence of arthritis. Almost 50% reported at least 1 nondisabling condition, and ∼20% reported at least 1 disabling condition. At followup, a greater proportion of individuals reporting worse SRH were older, of lower income, and lower education levels than those without worse SRH (Table 2).

Table 1. Prevalence of selected characteristics in the Canadian population at baseline, 1998/2000 Canadian longitudinal National Population Health Survey*
VariablePrevalence at baseline
  • *

    Values are percentage (95% confidence interval).

Self-rated health 
 Fair/poor9.1 (8.34–9.88)
 Good26.7 (25.55–27.84)
 Excellent/very good64.2 (62.98–65.42)
Usual intensity of pain/ discomfort 
 No pain85.8 (85.02–86.65)
 Mild4.4 (3.88–4.92)
 Moderate8.1 (7.42–8.73)
 Severe1.7 (1.42–1.96)
With activity limitation18.9 (17.90–19.98)
With arthritis17.0 (16.18–17.8)
≥ 1 non-disabling condition47.2 (45.91–48.43)
≥ 1 disabling condition19.0 (18.00–20.06)
≥ 1 life-threatening condition9.0 (8.30–9.67)
Table 2. Percentage distribution of self-rated health (SRH) status by sociodemographic characteristics in the Canadian population using data from the 1998/2000 Canadian longitudinal National Population Health Survey
VariableWorse SRHNo worse SRH
Age, years  
Income adequacy  
Education level  
 Less than secondary26.320.1
 Some post-secondary28.628.7

Almost 6% of the population reported developing arthritis, 20% reported worse SRH, ∼9% reported worse pain, and 7% developed an activity limitation. Other new chronic conditions were reported by 5–23% of respondents, with the greatest change being new nondisabling conditions. The proportions of respondents who developed each of the followup outcomes who also reported developing other outcomes are presented in Table 3. For example, almost 30% of individuals who developed arthritis reported worse SRH and 25% reported worse pain.

Table 3. Percentage of people who developed specified outcomes, and also reported developing additional outcomes, from the 1998/2000 Canadian longitudinal National Population Health Survey*
Developed outcomesOverall% with developed outcomes who developed additional outcomes
ArthritisWorse SRHWorse painActivity limitationNon-disabling conditionDisabling conditionLife-threatening condition
  • *

    SRH = self-rated health.

Worse SRH20.57.916.014.228.416.110.9
Worse pain8.516.538.824.336.325.511.6
Activity limitation7.011.441.829.637.026.512.4
Non-disabling condition23.09.625.313.411.217.98.3
Disabling condition10.611.031.120.317.438.79.0
Life-threatening condition5.

Results from the main effects models indicate that a significant relationship exists between arthritis onset and worse pain, developing an activity limitation, and worse SRH at followup (Table 4, Figure 1B). Respondents who reported developing arthritis were 30% more likely to report worse SRH and 3 times as likely to report worse pain, compared with those who did not develop arthritis. The development of new chronic conditions also had a substantial and independent effect on developing each of these outcomes. With each new life-threatening condition, the odds of reporting worse SRH more than doubled, and with the development of each new disabling condition, the odds of worse pain and new activity limitation more than doubled.

Table 4. Results from the main effects models for reporting specified outcome at followup, 1998/2000 Canadian longitudinal National Population Health Survey*
Variable of interestModel outcomes
Model 1 worse painModel 2 activity limitationModel 3 worse SRH
  • *

    Values are the odds ratios (95% confidence intervals) adjusted for age, sex, education, income, and all new conditions. SRH = self-rated health.

  • Reference groups are those people not developing specified conditions.

New arthritis3.09 (2.202–4.322)1.70 (1.176–2.420)1.30 (1.002–1.683)
New non-disabling condition1.27 (1.097–1.467)1.45 (1.231–1.697)1.19 (1.072–1.325)
New disabling condition2.23 (1.783–2.786)2.35 (1.852–2.981)1.52 (1.289–1.784)
New life-threatening condition1.51 (1.101–2.084)1.89 (1.355–2.638)2.35 (1.865–2.961)

The findings from the mediated models are illustrated in Table 5 and Figure 1C. With the introduction of worse pain into the model, arthritis onset no longer had a statistically significant association with worsening SRH. This is indicative of a mediating effect via worsening pain (Table 6). The relationship between arthritis onset and developing an activity limitation also became nonsignificant with the introduction of worse pain into the model, where a 55% decrease in effect was observed (Table 6). Again, this suggests that pain is a significant mediator of this relationship. A total decrease of almost 60% in arthritis-onset effect on worsening SRH was observed when both mediating factors were considered (Table 6). Although worsening pain and developing an activity limitation had independent associations with worsening SRH, the effect of worsening pain was diminished by 20% when developing an activity limitation was considered. This suggests a partially mediated effect via development of an activity limitation for worsening pain.

Table 5. Results from the mediated models for reporting specified outcome at followup, 1998/2000 Canadian longitudinal National Population Health Survey*
Variable of interestModel outcomes
Model 4 worse SRHModel 5 activity limitationModel 6 worse SRH
  • *

    Values are the odds ratios (95% confidence intervals) adjusted for age, sex, education, income, all new conditions and including worse pain for Models 4 and 5 and both worse pain and developing an activity limitation for Model 6. SRH = self-rated health.

  • Reference groups are those people not developing specified conditions.

New arthritis1.16 (0.886–1.509)1.31 (0.922–1.875)1.13 (0.865—1.485)
Worse pain2.15 (1.760–2.636)4.24 (3.308–5.437)1.90 (1.549–2.324)
Developed an activity limitation2.20 (1.797–2.693)
New non-disabling condition1.17 (1.050–1.303)1.41 (1.185–1.668)1.14 (1.024–1.275)
New disabling condition1.41 (1.197–1.659)2.06 (1.610–2.637)1.33 (1.129–1.565)
New life-threatening condition2.27 (1.794–2.870)1.74 (1.238–2.458)2.17 (1.730–2.732)
Table 6. Percentage of decrease in the effect of conditions on outcomes having accounted for worse pain and/or new activity limitation, 1998/2000 Canadian longitudinal National Population Health Survey*
Variable of interestOutcomes
NAL with variable worse pain (model 5 vs 2)Worse SRH with variable worse pain (model 4 vs 3)Worse SRH with variables worse pain and NAL (model 6 vs 3)Worse SRH with variables worse pain and NAL (model 6 vs 4)
  • *

    NAL = new activity limitation; SRH = self-rated health; vs = versus.

New arthritis55.146.756.718.8
New non-disabling8.910.526.317.6
New disabling21.521.236.519.5
New life-threatening16.95.913.37.9
Worse pain21.7

Tables 4 and 5 also present findings for the main effects of the onset of other chronic conditions. With the introduction of worsening pain and developing an activity limitation into the models, odds ratios were slightly reduced, suggesting some mediation via these variables. However, the extent of mediation was less than with the onset of arthritis. Percentage decreases in effects are presented in Table 6.


  1. Top of page
  2. Abstract
  7. Acknowledgements

In this analysis of longitudinal population survey data, the association between developing arthritis and worsening SRH was explained by the intermediary of worsening pain. The effect of this intermediary was partially mediated by development of an activity limitation. The very small residual direct effect of arthritis onset was not found to be significant. This is the first study to date that examines these relationships longitudinally and considers both of these hypothesized mediators. Worsening pain and developing activity limitations also explained a portion of the relationships between the onset of other chronic conditions and worsening SRH, but to a lesser extent than with arthritis onset. Pain seems to have the largest mediating effect on new disabling conditions. This may be related to the fact that back problems are the most frequent condition in this group. To some extent, the findings from this study explain why there appears to be a stronger relationship between pain and poor health status than between chronic diseases and poor health status (16, 17).

Most studies that have found a relationship between other chronic conditions and poorer self-reported health status have looked at these relationships either by examining a list of specific conditions or by considering the number of chronic conditions overall (16, 17, 23, 42, 43). This study examined chronic conditions by considering the onset frequency of broad types of conditions over the study period. Although a mediating role for pain was most evident with the onset of disabling conditions, there are likely other and differing mediators for different conditions. For example, we hypothesize that the relationship between the onset of some life-threatening conditions (such as diabetes and chronic heart disease) and SRH might be mediated by perceptions of disease seriousness in addition to the severity of actual physical symptoms. This is an area for further study.

In their studies examining the determinants of poor SRH in the elderly, including several conditions, Molarius and Janson (23) found pain and Hoeymans et al (42) found activity limitations to be significantly associated with poor SRH. Hoeymans et al documented a decreased effect of chronic conditions on poor SRH upon inclusion of activity limitations (42). In contrast, although Molarius and Janson conceptually acknowledged a role for pain, they did not report whether any change in effect of chronic conditions was observed when pain was considered (23). Also, neither study considered the effect of both pain and activity limitations.

This study has several limitations, one being that data were obtained through self report. However, it is known that the agreement between self-reported chronic conditions and medical records is fair to good, depending on the type of disease being investigated (47–49). Nonetheless, we contend that the ability or inability of medical records to corroborate a person's self report of having a condition, let alone whether they experience pain or have activity limitations, has little if any effect on how they report their own health status.

The grouping of conditions into broad categories is undoubtedly subjective. Also, any category that includes a condition(s) less severe than those remaining would inevitably bias results toward the null for the particular category. However, the significant associations with specified outcomes that we found by condition type are encouraging. As expected, we found life-threatening conditions to have the strongest association with worse SRH, and disabling conditions to have the strongest associations with worsening pain and developing an activity limitation.

Although the direction of the arrows as presented in Figure 1 is not supported by the current data and analyses, their direction is founded on theoretical and practical knowledge described earlier. Also, there is debate as to the proper method to test the statistical significance of mediation (50). Most of the work in this area has been based on continuous outcomes rather than on binary outcomes as considered in this study. However, despite the large sample size, the direct effect of arthritis onset, net of worsening pain, and development of activity limitations were found to be insignificant, leaving little doubt as to the significance of mediation.

Because poor SRH is predictive of several negative health outcomes, a better understanding of the relationship between disease and self-perceived health status is essential if target areas for prevention and intervention are sought. This study aimed to elucidate this relationship over time, particularly for developing arthritis. Worsening pain fully explained the effects of developing arthritis on worsening SRH; the effects of pain were further mediated by the development of activity limitations. This suggests that it was these symptoms rather than arthritis as such that were associated with worse SRH, at least in the short term, with implications for the clinical management of this condition. A better understanding of the nature of symptoms and other factors that mediate the effect of chronic conditions might point to important areas of intervention to reduce their impact on SRH.


  1. Top of page
  2. Abstract
  7. Acknowledgements

The authors thank Drs. Susan Bondy and Edward Adlaf for their input on data analysis and manuscript preparation.

  • 1

    The responsibility for the use and interpretation of the data is entirely that of the authors.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  • 1
    Statistical Commission and Economic Commission for Europe. Summary measures of population health and beyond: a look at US federal activities in measuring the health of populations. In: WHO Conference, Working Paper No. 11. Ottawa, Canada; 2000.
  • 2
    Hagerty MR, Cummins RA, Ferriss AL, Land K, Michalos AC, Peterson M, et al. Quality of life indexes for national policy: review and agenda for research. Soc Indic Res 2001; 55: 196.
  • 3
    Hennessy CH, Moriarty DG, Zack MM, Scherr PA, Brackbill R. Measuring health-related quality of life for public health surveillance. Public Health Rep 1994; 109: 66572.
  • 4
    Currey SS, Rao JK, Winfield JB, Callahan LF. Performance of a generic health-related quality of life measure in a clinic population with rheumatic disease. Arthritis Rheum 2003; 49: 65864.
  • 5
    Hurst DF, Boswell DL, Boogaard SE, Watson MW. The relationship of self-esteem to the health-related behaviors of the patients of a primary care clinic. Arch Fam Med 1997; 6: 6770.
  • 6
    Mulsant BH, Ganguli M, Seaberg EC. The relationship between self-rated health and depressive symptoms in an epidemiological sample of community-dwelling older adults. J Am Geriatr Soc 1997; 45: 9548.
  • 7
    Hillen T, Schaub R, Hiestermann A, Kirschner W, Robra BP. Self rating of health is associated with stressful life events, social support and residency in East and West Berlin shortly after the fall of the wall. J Epidemiol Community Health 2000; 54: 57580.
  • 8
    Idler EL, Kasl S. Health perceptions and survival: do global evaluations of health status really predict mortality? J Gerontol 1991; 46: S5565.
  • 9
    Branch L, Ku L. Transition probabilities to dependency, institutionalization and death among the elderly over a decade. J Aging Health 1989; 1: 370408.
  • 10
    Dominick KL, Ahern FM, Gold CH, Heller DA. Relationship of health-related quality of life to health care utilization and mortality among older adults. Aging Clin Exp Res 2002; 14: 499508.
  • 11
    Kennedy BS, Kasl SV, Vaccarino V. Repeated hospitalizations and self-rated health among the elderly: a multivariate failure time analysis. Am J Epidemiol 2001; 153: 23241.
  • 12
    Menec VH, Chipperfield JG. A prospective analysis of the relation between self-rated health and health care use among elderly Canadians. Can J Aging 2001; 20: 293306.
  • 13
    Idler EL, Benyamini Y. Self-rated health and mortality: a review of twenty-seven community studies. J Health Soc Behav 1997; 38: 2137.
  • 14
    Sorensen KH. State of health and its association with death among old people at three-years follow-up II. Consumption of medicine subjective and objective general health. Dan Med Bull 1988; 35: 597600.
  • 15
    Cott CA, Gignac MA, Badley EM. Determinants of self rated health for Canadians with chronic disease and disability. J Epidemiol Community Health 1999; 53: 7316.
  • 16
    Mantyselka PT, Turunen JH, Ahonen RS, Kumpusalo EA. Chronic pain and poor self-rated health. JAMA 2003; 290: 243542.
  • 17
    Reyes- Gibby CC, Aday L, Cleeland C. Impact of pain on self-rated health in the community-dwelling older adults. Pain 2002; 95: 7582.
  • 18
    Hirdes J, Forbes W. Factors associated with the maintenance of good self-rated health. J Aging Health 1993; 5: 10122.
  • 19
    House JS, Landis KR, Umberson D. Social relationships and health. Science 1988; 241: 5405.
  • 20
    Kaplan BH, Cassel JC, Gore S. Social support and health. Med Care 1977; 15 5 Suppl:47–58.
  • 21
    Stewart AL, Greenfield S, Hays RD, Wells K, Rogers WH, Berry SD, et al. Functional status and well-being of patients with chronic conditions: results from the Medical Outcomes Study [published erratum appears in JAMA 1989;262:2542]. JAMA 1989; 262: 90713.
  • 22
    Ferraro KF, Wilmoth JM. Measuring morbidity: disease counts, binary variables, and statistical power. J Gerontol B Psychol Sci Soc Sci 2000; 55: S17389.
  • 23
    Molarius A, Janson S. Self-rated health, chronic diseases, and symptoms among middle-aged and elderly men and women. J Clin Epidemiol 2002; 55: 36470.
  • 24
    Kaplan GA, Goldberg DE, Everson SA, Cohen RD, Salonen R, Tuomilehto J, et al. Perceived health status and morbidity and mortality: evidence from the Kuopio ischaemic heart disease risk factor study. Int J Epidemiol 1996; 25: 25965.
  • 25
    Manderbacka K, Lundberg O, Martikainen P. Do risk factors and health behaviours contribute to self-ratings of health? Soc Sci Med 1999; 48: 171320.
  • 26
    Ounpuu S, Chambers LW, Chan D, Yusuf S. Validity of the US Behavioral Risk Factor Surveillance System's health related quality of life survey tool in a group of older Canadians. Chronic Dis Can 2001; 22: 93101.
  • 27
    Krause NM, Jay GM. What do global self-rated health items measure? Med Care 1994; 32: 93042.
  • 28
    Andersson HI, Ejlertsson G, Leden I, Schersten B. Musculoskeletal chronic pain in general practice: studies of health care utilisation in comparison with pain prevalence. Scand J Prim Health Care 1999; 17: 8792.
  • 29
    Mantyselka P, Kumpusalo E, Ahonen R, Kumpusalo A, Kauhanen J, Viinamaki H, et al. Pain as a reason to visit the doctor: a study in Finnish primary health care. Pain 2001; 89: 17580.
  • 30
    Buskila D, Abramov G, Biton A, Neumann L. The prevalence of pain complaints in a general population in Israel and its implications for utilization of health services. J Rheumatol 2000; 27: 15215.
  • 31
    Badley EM, Rothman LM, Wang PP. Modeling physical dependence in arthritis: the relative contribution of specific disabilities and environmental factors. Arthritis Care Res 1998; 11: 33545.
  • 32
    Badley E. The impact of musculoskeletal disorders on the Canadian population. J Rheumatol 1992; 19: 33740.
  • 33
    Verbrugge LM, Lepkowski JM, Imanaka Y. Comorbidity and its impact on disability. Milbank Q 1989; 67: 45084.
  • 34
    Verbrugge LM, Lepkowski JM, Konkol LL. Levels of disability among US adults with arthritis. J Gerontol 1991; 46: S7183.
  • 35
    Verbrugge LM, Patrick DL. Seven chronic conditions: their impact on US adults' activity levels and use of medical services. Am J Public Health 1995; 85: 17382.
  • 36
    Badley EM, Rasooly I, Webster GK. Relative importance of musculoskeletal disorders as a cause of chronic health problems, disability, and health care utilization: findings from the 1990 Ontario Health Survey. J Rheumatol 1994; 21: 50514.
  • 37
    Lagace C, Perruccio A, Desmules M, Badley EM. The impact of arthritis on Canadians. In: BadleyEM, DesmulesM, editors. Arthritis in Canada: an ongoing challenge. Ottawa: Health Canada; 2003.
  • 38
    Picavet HS, Hazes JM. Prevalence of self reported musculoskeletal diseases is high. Ann Rheum Dis 2003; 62: 64450.
  • 39
    Wang G, Helmick CG, Macera C, Zhang P, Pratt M. Inactivity-associated medical costs among US adults with arthritis. Arthritis Rheum 2001; 45: 43945.
  • 40
    Public health and aging: projected prevalence of self-reported arthritis or chronic joint symptoms among persons aged >65 years: United States, 2005–2030. MMWR Morb Mortal Wkly Rep 2003; 52: 48991.
  • 41
    Raina P, Dukeshire S, Lindsay J, Chambers LW. Chronic conditions and disabilities among seniors: an analysis of population-based health and activity limitation surveys. Ann Epidemiol 1998; 8: 4029.
  • 42
    Hoeymans N, Feskens EJ, Kromhout D, van den Bos GA. The contribution of chronic conditions and disabilities to poor self-rated health in elderly men. J Gerontol A Biol Sci Med Sci 1999; 54: M5016.
  • 43
    Sprangers MA, de Regt EB, Andries F, van Agt HM, Bijl RV, de Boer JB, et al. Which chronic conditions are associated with better or poorer quality of life? J Clin Epidemiol 2000; 53: 895907.
  • 44
    Population Health Surveys Program. National population health survey cycle 4 (2000-2001), household component: longitudinal documentation. Ottawa, Canada: Statistics Canada; 2002.
  • 45
    Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51: 117382.
  • 46
    Aneshensel CS. Elaborating an explanation: antecedent, intervening and consequent variables. In: Theory-based data analysis for the social sciences. Thousand Oaks (CA): Pine Forge Press; 2002. p. 15890.
  • 47
    Beckett M, Weinstein M, Goldman N, Yu-Hsuan L. Do health interview surveys yield reliable data on chronic illness among older respondents? Am J Epidemiol 2000; 151: 31523.
  • 48
    Heliovaara M, Aromaa A, Klaukka T, Knekt P, Joukamaa M, Impivaara O. Reliability and validity of interview data on chronic diseases: the Mini-Finland Health Survey. J Clin Epidemiol 1993; 46: 18191.
  • 49
    Kriegsman DM, Penninx BW, van Eijk JT, Boeke AJ, Deeg DJ. Self-reports and general practitioner information on the presence of chronic diseases in community dwelling elderly: a study on the accuracy of patients' self-reports and on determinants of inaccuracy. J Clin Epidemiol 1996; 49: 140717.
  • 50
    MacKinnon DP, Lockwood CM, Hoffman JM, West SG, Sheets V. A comparison of methods to test mediation and other intervening variable effects. Psychol Methods 2002; 7: 83104.