SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Objective

To estimate the prevalence and co-occurrence of self-reported doctor-diagnosed arthritis, chronic joint symptoms (pain, aching, stiffness, or swelling on most days for a month), and transient joint symptoms (pain, aching, stiffness, or swelling but not on most days for a month), and to compare the sociodemographic characteristics, activity limitations, and health-related quality of life (HRQOL) of people with joint conditions with those who have no self-reported doctor-diagnosed arthritis and no joint symptoms.

Methods

Data from the 2004 population-based South Australian Health Omnibus Survey (n = 2,840, ages 18–96 years) were used in the study. Activity limitations were assessed using 10 activity limitations questions from the Short Form 36 health survey. HRQOL was assessed using the Assessment of Quality of Life scale.

Results

Half of all respondents reported having joint problems, with 26%, 11%, and 13% reporting self-reported doctor-diagnosed arthritis, chronic joint symptoms, and transient joint symptoms, respectively. Chronic joint conditions (self-reported doctor-diagnosed arthritis and chronic joint symptoms) accounted for 74% of all joint problems and were associated with higher odds of activity limitations and poorer HRQOL. The frequency of transient and chronic joint symptoms was highest among middle-aged participants (ages 45–54 years for transient and 45–64 years for chronic joint symptoms) and those who had a body mass index in the obese range. Prevalence of self-reported doctor-diagnosed arthritis increased with age and was higher among women and those who were overweight or obese.

Conclusion

This study documented the high prevalence and impact of joint conditions in the community. Chronic joint conditions affect daily life and are substantial barriers for effective public health interventions aimed at reducing obesity and inactivity.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Arthritis is a major public health problem, with estimated costs due to health care and lost productivity between 1–2% of the gross domestic product (1). At the individual level, people with arthritis incur higher health-related costs (2) and have worse quality of life than the general population (3). Monitoring the population's prevalence of arthritis is an important part of developing strategies to reduce the impact of this condition because it identifies population groups that may benefit from interventions.

In clinical practice, the diagnosis of arthritis is based on a combination of clinical assessment (history and physical examination) and radiographic and laboratory data (4–6). Given that detailed clinical investigations are costly and time consuming, alternative methods of identifying arthritis cases in population health surveys are still evolving. Prior to the mid 1990s, the identification of arthritis cases in population-based surveys was generally through self-report of one of the conditions classified as arthritis in the International Classification of Diseases, Ninth Revision (ICD-9) (7–10). However, this definition was found to be unreliable. In a survey of 472 rheumatology outpatients that assessed the concordance between ICD-9 based self-reported rheumatic conditions and diagnosis by physician, nearly one-third of respondents were unable to report their condition correctly (11).

As a result, guidelines by the US Centers for Disease Control and Prevention (CDC) recommended that the case definition of arthritis change from self-reported arthritis to self-report of having been diagnosed with arthritis by a health professional (12–14). Concurrently, the CDC introduced a question about the presence of chronic joint symptoms of pain, aching, and stiffness on most days for at least a month to try to capture those people with arthritis who had not consulted a health care professional. Those who had either self-reported doctor-diagnosed arthritis or chronic joint symptoms were classified as cases of arthritis. The approach of combining self-reported doctor-diagnosed arthritis and chronic joint symptoms, however, produced estimates of arthritis prevalence that have varied substantially over time (30% in 1996, 24% in 1997, and 33% in 2001) (15). Consequently, the CDC Arthritis Program issued an amended recommendation that self-reported doctor-diagnosed arthritis and chronic joint symptoms should not be combined into a single indicator, but that chronic joint symptoms should be used as an indicator of possible arthritis (16).

The prevalence of chronic joint symptoms has been estimated at 10% of the adult population (15, 16). However, the health service implications of identifying individuals with chronic joint symptoms are poorly understood because information on the impact of this condition on health, activity limitations, and health-related quality of life (HRQOL) is limited. It has also been reported that a substantial proportion of people who have chronic joint symptoms do not have arthritis on clinical examination (17, 18). These findings have given rise to concerns that people with transient joint symptoms (defined as the experience of joint pain, aching, stiffness, or swelling but not on most days for the past month), resulting from joint injury, for example, may be misclassified as cases of possible arthritis (19). Therefore, a better understanding of the prevalence and impact of both chronic and transient joint symptoms is warranted.

The aims of this study were to estimate the prevalence and co-occurrence of self-reported doctor-diagnosed arthritis, chronic joint symptoms, and transient joint symptoms, and to compare demographic and socioeconomic characteristics, activity limitations, and the HRQOL of individuals with joint problems (self-reported doctor-diagnosed arthritis or chronic or transient joint symptoms) with those who have no self-reported doctor-diagnosed arthritis and no joint symptoms.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The current study utilized data from the 2004 South Australian Health Omnibus Survey (20), which has been conducted annually since 1991 to collect health-related data used to plan, implement, and monitor health programs. The sample was selected using clustered multistage systematic area sampling, with information collected in face-to-face interviews. The interviews collect data on respondents' demographic characteristics in addition to health-related information. The participation rate was 66%, with 3,015 individuals ages 15–96 years taking part in the survey. Because the current study was concerned with the prevalence and impact of chronic joint symptoms in the adult population, the analyses were restricted to 2,890 persons ages ≥18 years; 125 individuals (4% of the total sample) ages <18 years were excluded from the analyses. To correct for nonresponse, the survey data were weighted by the inverse probability of selection within the collection district and by estimated population benchmarks for age, sex, and geographic region. The weighted sample size was 2,840 persons. Further details of the survey have been reported elsewhere (20).

The study was approved by the Human Research Ethics Committee of the South Australian Department of Health Research and was carried out in compliance with the Helsinki Declaration.

Joint symptoms and arthritis.

Two questions were used to assess the presence and duration of joint symptoms (excluding back or neck problems): “Do you have pain, aching, stiffness, or swelling in or around a joint (yes/no)?” and “Were the symptoms present on most days for at least a month (yes/no)?” Arthritis was defined as self-reported doctor-diagnosed arthritis and was assessed with the question, “Have you ever been told by a doctor that you have arthritis (yes/no)?” Based on their responses to these questions, participants were classified into 1 of 4 mutually exclusive groups: those who answered no to the presence of joint symptoms and no to the arthritis question were classified as no joint problems; those who answered yes to the self-reported doctor-diagnosed arthritis question were classified as arthritis, regardless of their answers to the joint symptoms questions; those who answered yes to the presence of joint symptoms but no to whether they had been present on most days for at least a month, and no to the arthritis question, were classified as transient joint symptoms; and those who answered yes to the presence of joint symptoms, yes that symptoms had been present on most days for at least a month, and no to the arthritis question were classified as chronic joint symptoms.

Demographic and socioeconomic factors.

Information was collected on the participants' age, sex, residence (rural versus metropolitan), education, occupation, and income. Age was categorized into <35 years, 35–44 years, 45–54 years, 55–64 years, 65–74 years, and >74 years. Education was coded as primary (<7 years), secondary (7–12 years), or tertiary (completed further education, such as technical college or university). Occupation was determined by asking the participants, “What kind of work have you done for most of your life?” Responses were recorded verbatim and categorized into standard groupings (21): professional (managerial, professional, and paraprofessional occupations) and nonprofessional (trades, sales, unskilled labor, voluntary workers, and home duty). Welfare benefit recipients and students were categorized as an “other” occupational group. Income was assessed as “income before tax from all sources over the past 12 months,” and was grouped into the categories <$20,000, $20,000–40,000, $41,000–60,000, and >$60,000.

Body mass index (BMI).

BMI was calculated by dividing the participants' self-reported weight in kilograms by the square of their self-reported height in meters. BMI values were classified into the categories normal weight (<25 kg/m2), overweight (≥25 and <30 kg/m2), and obese (≥30 kg/m2) (22).

Limitations on activity.

Ten activity-related questions from the Short Form 36 health survey (23) were used to assess activity limitations. The questions were preceded by the statement, “The following questions are about activities you might do during a typical day. Does your health now limit you in these activities? If so, how much?” The following activities were assessed: vigorous activities; moderate activities; lifting or carrying groceries; climbing several flights of stairs; climbing 1 flight of stairs; bending, kneeling, or stooping; walking more than a kilometer; walking several hundred meters; walking 100 meters; and bathing or getting dressed.

Responses to these questions were scored on a 3-point scale (where 1 = limited a lot, 2 = limited a little, and 3 = not limited at all). For ease of interpretation of the results, the limited a lot and limited a little options were collapsed into one category in the analyses, with the final response categories being 1 (limited) and 0 (not limited).

HRQOL.

The Assessment of Quality of Life scale (AQOL) was used to measure HRQOL. The AQOL is a generic HRQOL tool that has 12 items in 4 subscales, including independent living, social relationships, physical senses, and psychological well-being (24). The subscales are combined into an overall utility score ranging from −0.04 (worst possible HRQOL) to 1.00 (full HRQOL). The subscale weights were determined using a time-tradeoff method in a general population sample (24). The instrument has been effectively used with a variety of health conditions, including arthritis (25, 26).

Statistical analyses.

Results were analyzed using SPSS, version 15.0.0 (27), with estimates weighted by age, sex, and demographic region (20). The nominal level of statistical significance (alpha) was set at 0.05 (corresponding to 95% confidence intervals [95% CIs]), unless otherwise specified.

The prevalence of self-reported doctor-diagnosed arthritis, chronic joint symptoms, and transient joint symptoms was determined using proportions. Chi-square tests were used to identify demographic and socioeconomic variables associated with arthritis/joint symptom status. Multinomial logistic regression, with the group who had no joint problems as a reference, was used to assess demographic and socioeconomic characteristics of the arthritis/joint symptom groups. Variables that were significantly associated with arthritis/joint symptom status in univariate chi-square analyses were entered into the model simultaneously. Reference categories for the sociodemographic variables included male sex, age <35 years, rural residence, primary level of education, professionals, income <$20,000 per annum, and BMI <25 kg/m2 (normal range).

Logistic regression was used to assess the associations between activity limitations and arthritis/joint symptom status, after adjusting for age, sex, and BMI. The associations were assessed using odds ratios (ORs) with 99% CIs. We used 99% CIs to correct for the increased risk of Type I error due to the multiple tests.

Analysis of covariance (ANCOVA) was used to compare the HRQOL scores of the arthritis/joint symptom groups, after adjusting for age and sex. AQOL scores were used as the dependent variable, arthritis/joint symptom status as the independent variable, and age and sex as covariates. Results of the assumptions testing showed that homogeneity of variance was violated. Therefore, the alpha level of statistical significance was reduced to 0.01 to allow for a more stringent test of the differences in the average HRQOL score of the arthritis/joint symptom groups (28). To assess the sensitivity of the ANCOVA model to the violation of the homoscedasticity assumption, AQOL scores were cube-transformed to produce a more homoscedastic distribution, and the analysis was repeated using the transformed variable. Group differences in HRQOL were assessed using mean AQOL scores with 99% CIs.

Missing data were minimal, but were highest for income (9.0%), BMI (8.1%), and occupation status (3.4%). To maximize the data available for the analysis and reduce the possibility of bias due to missing data, participants who had missing values for income, occupation, or BMI were assigned a dummy category for the analyses.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Participants.

Sociodemographic and BMI characteristics of the survey participants are presented in Table 1. The participants were representative of the Australian population and ranged in age from 18–96 years (median 46 years), with 49.2% being men. Whereas just over half (51.6%) of participants had a tertiary level of education, a large proportion (67.2%) were employed in nonprofessional occupations. Half (50.4%) of the participants had a BMI in the obese or overweight range.

Table 1. Sociodemographic characteristics and body mass index of study participants (n = 2,840)*
 MenWomenTotal
  • *

    Values are the number (percentage) of participants.

Age, years   
 <35431 (30.9)409 (28.3)840 (29.6)
 35–44277 (19.8)272 (18.9)549 (19.3)
 45–54256 (18.3)259 (17.9)515 (18.1)
 55–64197 (14.1)198 (13.7)395 (13.9)
 ≥65236 (16.9)305 (21.2)541 (19.0)
Area of state   
 Rural433 (31.0)410 (28.4)843 (29.7)
 Metropolitan964 (69.0)1,033 (71.6)1,997 (70.3)
Education level   
 Primary169 (12.1)284 (19.7)453 (15.9)
 Secondary395 (28.2)528 (36.6)923 (32.5)
 Tertiary834 (59.7)630 (43.7)1,464 (51.6)
Occupation   
 Nonprofessional888 (63.6)1,021 (70.7)1,909 (67.2)
 Professional418 (29.9)363 (25.2)781 (27.5)
 Other23 (1.7)30 (2.0)53 (1.9)
 Missing68 (4.9)29 (2.0)98 (3.4)
Income   
 <$20,000235 (16.8)365 (25.3)601 (21.1)
 $20,000–40,000251 (18.0)300 (20.8)551 (19.4)
 $41,000–60,000243 (17.4)228 (15.8)471 (16.6)
 >$60,000555 (39.8)408 (28.2)963 (33.9)
 Missing112 (8.0)142 (9.9)255 (9.0)
Body mass index   
 Normal523 (37.5)656 (45.4)1,179 (41.5)
 Overweight554 (39.6)361 (25.0)915 (32.2)
 Obese244 (17.5)272 (18.9)516 (18.2)
 Missing77 (5.5)154 (10.7)231 (8.1)

Prevalence of joint symptoms and arthritis.

The frequencies of self-reported doctor-diagnosed arthritis, chronic joint symptoms, transient joint symptoms, and their co-occurrence are presented in Figure 1. Half of all respondents (49.8%) reported experiencing ≥1 type of joint problem, including self-reported doctor-diagnosed arthritis, chronic joint symptoms, or transient joint symptoms. Among those who reported joint problems, chronic conditions (self-reported doctor-diagnosed arthritis or chronic joint symptoms) were more frequent than transient joint problems, accounting for 74% of all joint problems. Overall, 26.0% of participants had self-reported doctor-diagnosed arthritis, 23.8% had chronic joint symptoms, and 17.9% had transient joint symptoms. Self-reported doctor-diagnosed arthritis and chronic joint symptoms co-occurred in 13.2% of participants, 4.7% had self-reported doctor-diagnosed arthritis with transient joint symptoms, and 8.1% reported arthritis without joint symptoms. The prevalences of chronic joint symptoms alone (without self-reported doctor-diagnosed arthritis) and transient joint symptoms alone were 10.7% and 13.1%, respectively. Among women, 11.5% had transient joint symptoms alone, 10.6% had chronic joint symptoms alone, and 30.0% had self-reported doctor-diagnosed arthritis. Among men, 14.8% had transient joint symptoms alone, 10.8% had chronic joint symptoms alone, and 21.9% had self-reported doctor-diagnosed arthritis.

thumbnail image

Figure 1. Venn diagram of the frequency of transient joint symptoms, chronic joint symptoms, self-reported doctor-diagnosed arthritis, and their co-occurrence.

Download figure to PowerPoint

Characteristics of the joint symptoms/arthritis groups.

The prevalence of chronic and transient joint symptoms peaked between 35–54 years of age, followed by a decline among the older age groups (Table 2). Self-reported doctor-diagnosed arthritis was uncommon in participants age <35 years and increased monotonically with age. The association between age and arthritis/joint symptom status was significant: χ2(12) = 460.1, P < 0.001. In addition to age, arthritis/joint symptom status was associated with sex, education, occupation, income, and BMI, but not with the area of residence (Table 2).

Table 2. Prevalence of joint symptoms and self-reported doctor-diagnosed arthritis by socioeconomic status and body mass index (n = 2,840)*
 No joint problemsTransient joint symptomsChronic joint symptomsSelf-reported doctor-diagnosed arthritisChi-square test
  • *

    Values are the number (percentage) of participants unless otherwise indicated.

Age, years    χ2(12) = 460.1, P < 0.001
 <35573 (68.2)121 (14.5)93 (11.0)53 (6.3) 
 35–44313 (57.0)85 (15.5)59 (10.8)91 (16.6) 
 45–54220 (42.7)81 (15.8)62 (12.0)152 (29.5) 
 55–64148 (37.4)42 (10.6)47 (11.8)159 (40.3) 
 ≥65172 (31.8)43 (7.9)43 (7.9)283 (52.3) 
Sex    χ2(3) = 26.8, P < 0.001
 Male734 (52.5)207 (14.8)150 (10.8)306 (21.9) 
 Female692 (47.9)166 (11.5)153 (10.6)433 (30.0) 
Area of residence    χ2(3) = 1.9, P = 0.590
 Rural432 (51.2)107 (12.7)81 (9.6)224 (26.5) 
 Metropolitan994 (49.8)266 (13.3)223 (11.2)515 (25.8) 
Education level    χ2(6) = 117.0, P < 0.001
 Primary157 (34.6)38 (8.4)52 (11.4)206 (45.5) 
 Secondary508 (55.0)130 (14.1)93 (10.0)192 (20.8) 
 Tertiary761 (51.9)205 (14.0)159 (10.9)340 (23.2) 
Occupation    χ2(9) = 26.8, P = 0.002
 Nonprofessional956 (50.1)241 (12.6)192 (10.0)520 (27.2) 
 Professional383 (49.0)105 (13.4)94 (12.0)200 (25.6) 
 Other28 (54.0)6 (12.2)11 (21.4)7 (12.4) 
 Missing59 (59.9)20 (20.5)7 (6.9)12 (12.7) 
Income    χ2(12) = 152.4, P < 0.001
 <$20,000231 (38.5)54 (9.0)51 (8.4)265 (44.0) 
 $20,000–40,000266 (48.3)79 (14.4)57 (10.3)149 (27.0) 
 $41,000–60,000265 (56.2)65 (13.8)56 (12.0)85 (18.0) 
 >$60,000528 (54.9)148 (15.4)110 (11.5)176 (18.3) 
 Missing135 (53.2)26 (10.2)29 (11.4)64 (25.2) 
Body mass index    χ2(9) = 61.8, P < 0.001
 Normal664 (56.3)157 (13.3)107 (9.1)251 (21.3) 
 Overweight450 (49.2)123 (13.4)93 (10.2)249 (27.3) 
 Obese196 (38.0)73 (14.2)78 (15.2)168 (32.6) 
 Missing116 (50.3)20 (8.7)25 (10.7)70 (30.3) 

Results showed that age, sex, and BMI contributed to the odds of reporting arthritis/joint symptoms after adjusting for education, occupation, and income (Table 3). The proportion of transient joint symptoms was highest among people ages 45–54 years and those who had BMI in the obese range. Higher frequency of chronic joint symptoms was associated with being 45–64 years old and with having a BMI in the obese range. The prevalence of self-reported doctor-diagnosed arthritis increased with age and was higher for women, and among those who had a BMI in the overweight or obese range. People with an income between $41,000–60,000 per annum had a slightly lower frequency of self-reported doctor-diagnosed arthritis than those with an income <$20,000 (OR 0.69, 95% CI 0.49–0.98), although income did not contribute significantly to the model (Table 3).

Table 3. Joint symptoms and self-reported doctor-diagnosed arthritis across the population (n = 2,840) groups*
 Transient joint symptomsChronic joint symptomsSelf-reported doctor- diagnosed arthritisLikelihood ratio test
  • *

    Values are the odds ratio (95% confidence interval) unless otherwise indicated, and are derived from a multinomial logistic regression analysis using joint symptoms/arthritis status as the dependent variable and age, sex, education, occupation, income, and body mass index as predictors and no joint symptoms as the reference group.

  • Reference category.

Age   χ2(12) = 299.84, P < 0.001
 <35 years1.001.001.00 
 35–44 years1.21 (0.88–1.65)1.04 (0.72–1.49)2.86 (1.98–4.15) 
 45–54 years1.68 (1.21–2.34)1.50 (1.04–2.17)6.60 (4.62–9.42) 
 55–64 years1.33 (0.88–2.00)1.71 (1.13–2.60)9.88 (6.79–14.37) 
 ≥65 years1.28 (0.82–1.99)1.45 (0.91–2.30)14.11 (9.66–20.61) 
Sex   χ2(3) = 17.32, P = 0.001
 Male1.001.001.00 
 Female0.89 (0.70–1.13)1.10 (0.85–1.43)1.46 (1.19–1.79) 
Education   χ2(6) = 7.70, P = 0.261
 Primary1.001.001.00 
 Secondary1.11 (0.72–1.72)0.62 (0.41–0.95)0.77 (0.57–1.04) 
 Tertiary1.11 (0.73–1.70)0.69 (0.46–1.04)0.88 (0.66–1.18) 
Occupation   χ2(9) = 11.16, P = 0.265
 Nonprofessional1.001.001.00 
 Professional1.08 (0.83–1.40)1.24 (0.94–1.63)1.04 (0.83–1.30) 
 Other0.85 (0.35–2.04)1.91 (0.93–3.92)0.58 (0.24–1.42) 
 Missing1.43 (0.83–2.45)0.70 (0.31–1.59)1.09 (0.56–2.12) 
Income   χ2(12) = 14.28, P = 0.283
 <$20,0001.001.001.00 
 $20,000–40,0001.29 (0.86–1.93)1.17 (0.76–1.81)0.78 (0.58–1.06) 
 $41,000–60,0001.07 (0.70–1.66)1.32 (0.83–2.09)0.69 (0.49–0.98) 
 >$60,0001.14 (0.77–1.70)1.29 (0.84–1.99)0.78 (0.57–1.06) 
 Missing0.87 (0.51–1.47)1.22 (0.72–2.06)0.71 (0.48–1.04) 
Body mass index   χ2(9) = 46.33, P < 0.001
 Normal1.001.001.00 
 Overweight1.08 (0.82–1.41)1.21 (0.89–1.65)1.33 (1.05–1.68) 
 Obese1.50 (1.08–2.07)2.33 (1.66–3.27)2.12 (1.61–2.80) 
 Missing0.78 (0.47–1.29)1.30 (0.80–2.13)1.63 (1.12–2.37) 

Further multiple logistic regression analyses (data not shown), used to compare the sociodemographic characteristics of the arthritis/joint symptom groups, showed that joint symptom status (transient versus chronic) was associated with education and BMI, but not with age, sex, or income. People with chronic joint symptoms were 80% more likely to have only a primary level of education (OR 1.79, 95% CI 1.08–2.97) and 60% more likely to have a BMI in the obese range (OR 1.62, 95% CI 1.06–2.48) than the transient joint symptoms group. Comparison of the chronic joint symptoms and self-reported doctor-diagnosed arthritis groups showed that the members of the latter group were 5 times less likely to be ages <35 years (OR 0.21, 95% CI 0.14–0.31) and 60% more likely to have an annual income of <$20,000 (OR 1.62, 95% CI 1.09–2.40) than people with chronic joint symptoms.

Activity limitations.

The majority of participants (65.7%) reported being limited in ≥1 of the 10 activities assessed. The prevalence of activity limitations was generally lowest for those who had no joint problems, followed by the transient and chronic joint symptoms groups, with the highest prevalence of limitations among those who had self-reported doctor-diagnosed arthritis. Across all groups, the limitations were reported the most frequently for vigorous activities (in 46.8%, 55.5%, 68.2%, and 83.4% of participants with no joint problems, transient joint symptoms, chronic joint symptoms, and self-reported doctor-diagnosed arthritis, respectively) and the least frequently for bathing or getting dressed (in 1.6%, 2.7%, 3.5%, and 9.6%, respectively). Overall, women reported more limitations (mean 2.72, 99% CI 2.52–2.92) than men (mean 1.82, 99% CI 1.66–1.99). Pearson's product-moment correlation coefficient (r) also showed that the number of activity limitations increased with age (r = 0.49, P < 0.001) and BMI (r = 0.17, P < 0.001).

From the 10 activities assessed, people with transient joint symptoms were more limited than those who had no joint problems only in bending, kneeling, or stooping (OR 1.79, 99% CI 1.24–2.59), even after adjusting for age, sex, and BMI (Table 4). Those who had chronic joint symptoms generally had an ∼2-fold increase in the odds of being limited in all activities, except bathing or getting dressed, than people with no joint problems. The highest odds of having limitations for the chronic joint symptoms group was in walking several hundred meters (OR 3.29, 99% CI 1.90–5.70) and in climbing 1 flight of stairs (OR 2.77, 99% CI 1.56–4.92). Compared with those who had no joint problems, the participants in the chronic joint symptoms group were least likely to be limited in climbing several flights of stairs (OR 2.03, 99% CI 1.35–3.05) and in moderate activities (OR 2.00, 99% CI 1.28–3.12). The participants in the self-reported doctor-diagnosed arthritis group were much more likely to be limited than those who had no joint problems in all types of activities, with their likelihood of limitations generally being 3-fold or higher. The highest odds of having limitations for the participants in the self-reported doctor-diagnosed arthritis group were in bathing or getting dressed (OR 4.24, 99% CI 2.11–8.55) and in bending, kneeling, or stooping (OR 4.02, 99% CI 2.99–5.40). The lowest likelihood of limitations for this group was in walking more than a kilometer (OR 2.65, 99% CI 1.91–3.67) and in climbing several flights of stairs (OR 2.61, 99% CI 1.92–3.53).

Table 4. Activity limitations across arthritis/joint symptoms groups (n = 2,840)*
 Transient joint symptomsChronic joint symptomsSelf-reported doctor- diagnosed arthritis
  • *

    Values are the odds ratio (99% confidence interval) and derived from logistic regression analyses using activity limitations (yes/no) questions as the dependent variables, joint symptoms/arthritis status as the predictor, and age, sex, and body mass index as the covariates. Reference group was no joint problems.

Vigorous activities1.31 (0.94–1.83)2.18 (1.48–3.21)3.08 (2.21–4.29)
Moderate activities0.83 (0.50–1.37)2.00 (1.28–3.12)3.19 (2.31–4.40)
Lifting or carrying groceries0.80 (0.45–1.41)2.20 (1.37–3.54)3.18 (2.24–4.50)
Climbing several flights of stairs1.19 (0.79–1.79)2.03 (1.35–3.05)2.61 (1.92–3.53)
Climbing 1 flight of stairs1.12 (0.57–2.21)2.77 (1.56–4.92)3.80 (2.52–5.74)
Bending, kneeling, or stooping1.79 (1.24–2.59)2.58 (1.76–3.79)4.02 (2.99–5.40)
Walking >1 kilometer1.06 (0.66–1.72)2.40 (1.55–3.71)2.65 (1.91–3.67)
Walking several hundred meters0.83 (0.40–1.73)3.29 (1.90–5.70)3.42 (2.26–5.16)
Walking 100 meters0.83 (0.31–2.22)2.41 (1.15–5.03)3.66 (2.18–6.14)
Bathing or dressing yourself1.64 (0.58–4.63)2.37 (0.89–6.28)4.24 (2.11–8.55)

HRQOL scores.

HRQOL was positively skewed across all arthritis/joint symptom groups, with the majority in each group reporting high levels of HRQOL (Figure 2). However, whereas nearly two-thirds of people with no joint problems (71%) or with transient joint symptoms (67%) had an HRQOL at the population norm or higher (AQOL utility score 0.83) (29), more than half of people (54%) with chronic joint symptoms and two-thirds of people (62%) with self-reported doctor-diagnosed arthritis had an HRQOL below the population norm.

thumbnail image

Figure 2. Distribution of health-related quality of life (by Assessment of Quality of Life [AQOL] scale utility score) for the arthritis/joint symptoms groups. Shaded boxes represent individuals with the middle 50% of scores, solid lines inside the boxes represent the median scores, top whiskers represent the highest 25% of scores, and bottom whiskers represent the lowest 25% of scores. Circles beyond the bottom whiskers represent outliers (values that are between 1.5 and 3 times the interquartile range [IQR]), and stars represent extreme values (>3 times the IQR). The IQR is the difference between the 75th and 25th percentiles and corresponds to the length of the box (27).

Download figure to PowerPoint

Pearson's product-moment correlation coefficient showed that HRQOL was virtually unrelated to BMI (r = −0.09, P < 0.001), but declined slightly with age (r = −0.25, P < 0.001). Average HRQOL was also marginally lower in women (mean 0.79, 99% CI 0.78–0.81) than in men (mean 0.82, 99% CI 0.81–0.84). After adjusting for age and sex, HRQOL differed significantly between arthritis/joint symptom groups (F[3,2920] = 80.88, P < 0.001). People with no joint problems had the best HRQOL (mean 0.85, 99% CI 0.84–0.87), followed by the transient joint symptoms group (mean 0.84, 99% CI 0.81–0.86), the chronic joint symptoms group (mean 0.75, 99% CI 0.72–0.78), and the self-reported doctor-diagnosed arthritis group (mean 0.72, 99% CI 0.70–0.74). People with transient joint symptoms did not differ from those who had no joint problems or from the Australian population norm on HRQOL.

ANCOVA models conducted with untransformed and cube-transformed AQOL scores produced equivalent results, indicating that the heteroscedasticity had minimal influence on the conclusions of this study.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Using the 2004 population-based South Australian Health Omnibus Survey, we determined both the prevalence and impact of self-reported doctor-diagnosed arthritis, chronic joint symptoms, and transient joint symptoms in the general population. The prevalence of joint problems was high, with 50% of participants reporting either joint symptoms or self-reported doctor-diagnosed arthritis. Of the respondents, 11% had chronic joint symptoms, 13% had transient joint symptoms, and approximately one-quarter (26%) reported having a prior diagnosis of arthritis. Chronic joint conditions (chronic joint symptoms and self-reported doctor-diagnosed arthritis) accounted for 74% of all joint problems. Increased likelihood of transient joint symptoms was associated with being 45–54 years old and having a BMI in the obese range. The presence of chronic joint symptoms was associated with being 45–64 years old and having a BMI in the obese range. The frequency of self-reported doctor-diagnosed arthritis greatly increased with age and was higher among women and those who had a BMI in the overweight or obese range. Although on average transient joint symptoms were not associated with poor HRQOL, people with transient joint symptoms were more limited in bending, kneeling, or stooping than those who had no joint problems. People who had either chronic joint symptoms or arthritis were more likely to have activity limitations and worse HRQOL than people with no joint problems or people with transient joint symptoms. The self-reported doctor-diagnosed arthritis group experienced the most activity limitations and the poorest HRQOL. These findings address an important gap in previous research by providing information on the population groups with various types of joint problems and the consequences of each on daily function and HRQOL.

Our prevalence estimates of 13% for transient joint symptoms, 11% for chronic joint symptoms, and 26% for self-reported doctor-diagnosed arthritis among Australian adults interviewed face-to-face are similar to the results from the US Behavioral Risk Factors Surveillance System. Using a similar set of screening questions in a telephone interview, this survey estimated that 12%, 10%, and 23% of US adults have transient joint symptoms, chronic joint symptoms, and self-reported doctor-diagnosed arthritis, respectively (30). Although the characteristics of people with joint pain have been previously assessed in several studies (31–33), only a few studies simultaneously examined the distribution of chronic joint symptoms and self-reported doctor-diagnosed arthritis across population groups (16, 30, 34), with conflicting results. Virtually all such studies were conducted in the US, with consistent reports that increased prevalence of self-reported doctor-diagnosed arthritis was associated with older age, female sex, and low socioeconomic status. However, findings about the characteristics of people with chronic joint symptoms are less consistent. In some investigations, people with chronic joint symptoms did not differ from those who had no joint problems on age, sex, or educational attainment (30, 34), whereas in other investigations, higher frequency of chronic joint symptoms was associated with being <65 years of age and male (16). In our study, chronic joint symptoms were most frequent among people ages 45–64 years old, after controlling for sex, area of residence, education, occupation, income, and BMI. Direct comparisons of the chronic joint symptoms group with those who had transient joint symptoms and self-reported doctordiagnosed arthritis also showed that people with chronic joint symptoms had less education and higher BMI than people with transient joint symptoms, and were generally younger and had higher incomes than those who had self-reported doctor-diagnosed arthritis.

However, it is important to note that our study is based on a probability sample of Australian households, with data collected in face-to-face interviews. One of the previous US-based studies (34) was conducted with a probability sample of 851 Chicago-area residents ages ≥45 years. Other studies were based on probability samples of the US general population ages ≥18 years (16, 30) with data collected in telephone interviews. Because different sampling and data collection methods systematically exclude certain population groups (for example, the homeless, and people without a telephone), differences in the findings between our study and past studies may reflect the study designs rather than real differences in the distribution of chronic joint symptoms between population groups in Australia and the US.

Previous studies have identified high BMI as a risk factor for arthritis and chronic joint pain (30, 33, 35). The results of our study extend previous findings by showing that having a BMI in the obese range is associated with increased likelihood of transient joint symptoms, even after controlling for the effects of age, sex, and socioeconomic status. In addition, people with a BMI in the obese range may be more likely to have chronic joint symptoms rather than transient joint symptoms. Although the cross-sectional design of this study does not allow conclusions about the causal relationship between BMI and joint problems, longitudinal evidence suggests that people with excess weight are at a higher risk for arthritis later in life (36, 37).

A substantial body of literature documents the high personal burden of arthritis, with adverse impacts on health and functioning (38–40). Chronic musculoskeletal pain is also associated with disturbed sleep (41) and depression (42). The high morbidity among people with chronic joint symptoms in our study further suggests that this population group may benefit from intervention programs directed at preventing additional disability. Our results also indicate the progressive increase in activity limitations and the decline in HRQOL among the joint conditions studied, beginning with transient joint symptoms, chronic joint symptoms, and self-reported doctor-diagnosed arthritis. This suggests that these groups may represent a severity continuum of underlying joint pathology. Our finding that the prevalences of both transient and chronic joint symptoms decrease with age, whereas the prevalence of self-reported doctor-diagnosed arthritis increases with age, provides some support for this hypothesis. However, longitudinal studies are needed to explore the possibility of conversion from transient to chronic joint symptoms and from chronic joint symptoms to arthritis.

The risk of activity limitations in chronic musculoskeletal conditions is well documented (8, 43, 44). Several studies have also linked functional impairment with increased BMI (45, 46), with evidence to suggest that a high BMI predates the development of activity limitations in people with joint pain (40). A recent meta-analysis of randomized controlled trials also reported a decrease in physical disability among overweight patients with knee osteoarthritis after a weight reduction program (47). Exercise and weight loss have therefore become an important part of arthritis treatment and management (48, 49). However, our study also shows that the presence of joint problems is a strong risk factor for activity impairment independent of BMI. This finding has important implications for arthritis management. Similarly, although weight loss and exercise programs for people with a range of chronic conditions can improve overall health status and reduce the risks of heart disease and diabetes mellitus, mobility limitations experienced by those with concurrent joint problems, including difficulties with walking, bending, and climbing stairs, need to be taken into consideration when designing and implementing these programs.

Information on arthritis prevalence, with detailed information on the specific limitations that people with this condition experience, highlights the extent and nature of the arthritis burden, and supports health service program planning and the design of interventions to prevent and reduce arthritis-related disability. However, a limitation of our study was the lack of clinical assessment to verify self-reported arthritis diagnosis and to determine the underlying pathology of joint symptoms. Previous studies have shown that the estimates of the prevalences of painful musculoskeletal conditions are greatly influenced by the method of case definition (42, 50). The validity of chronic joint symptoms as an indicator of possible arthritis has been examined in a limited number of studies (17, 18), with results suggesting that only ∼60% of people ages 45–64 years who meet the case definition for chronic joint symptoms have some form of arthritis on clinical examination. The accuracy of chronic joint symptoms as an indicator of arthritis was reported to increase with age, with 90% of individuals with chronic joint symptoms having clinically confirmed arthritis at ages ≥65 years. Although no comparable data are available for individuals ages <45 years, this suggests that using chronic joint symptoms as an indicator of arthritis overestimates its prevalence, especially among younger individuals.

Therefore, to better understand the implications of identifying individuals with chronic joint symptoms for health care services, future studies are needed to investigate the relevant clinical outcomes and sequela of chronic joint symptoms. Despite the uncertainty about pathology of chronic joint symptoms, the results of this study also suggest that this indicator identifies individuals with compromised health and well-being, therefore providing potentially valuable information for health care planners. A more detailed understanding of people's behavior and capacity to engage in exercise programs while living with chronic joint conditions will support the development of more focused primary and secondary prevention programs.

A limitation of this study was a relatively low response rate (66%), which could potentially jeopardize the external validity of the results. However, because the survey data were weighted by age, sex, and geographic area, the impact of nonresponse on generalizability is expected to be minimal.

The results of this study document the high prevalence and impact of joint conditions in the community. Although clinical investigations are required to confirm the validity of chronic joint symptoms as an indicator of possible arthritis, the chronic joint conditions studied here are substantial barriers to effective public health interventions aimed at reducing primary and secondary risk factors such as obesity and inactivity. With the aging population, increasing obesity rates, and declining physical activity, arthritis prevalence is likely to increase and impede individuals' ability to participate in the work force, fulfill career roles, and be active members of the community more generally.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Ms Busija 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 conception and design. Busija, Buchbinder, Osborne.

Acquisition of data. Osborne.

Analysis and interpretation of data. Busija, Buchbinder, Osborne.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The authors wish to thank Anne Taylor from the South Australian Department of Health for providing access to the data and commenting on drafts.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  • 1
    March LM, Bachmeier CJ. Economics of osteoarthritis: a global perspective. Baillieres Clin Rheumatol 1997; 11: 81734.
  • 2
    Dunlop DD, Manheim LM, Yelin EH, Song J, Chang RW. The costs of arthritis. Arthritis Rheum 2003; 49: 10113.
  • 3
    Ackerman IN, Graves SE, Wicks IP, Bennell KL, Osborne RH. Severely compromised quality of life in women and those of lower socioeconomic status waiting for joint replacement surgery. Arthritis Rheum 2005; 53: 6538.
  • 4
    Altman RD. Criteria for the classification of osteoarthritis of the knee and hip. Scand J Rheumatol Suppl 1987; 65: 319.
  • 5
    Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, Brandt K, et al. The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hip. Arthritis Rheum 1991; 34: 50514.
  • 6
    Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, Brandt K, et al. The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hand. Arthritis Rheum 1990; 33: 160110.
  • 7
    Centers for Disease Control and Prevention. Progress in chronic disease prevention prevalence of arthritic conditions: United States, 1987. MMWR CDC Surveill Summ 1990; 39: 99100, 102.
  • 8
    Centers for Disease Control and Prevention. Arthritis prevalence and activity limitations: 1990. Wkly Epidemiol Rec 1994; 69: 2847.
  • 9
    Centers for Disease Control and Prevention. Prevalence and impact of arthritis among women: United States, 1989–1991. JAMA 1995; 273: 18202.
  • 10
    Centers for Disease Control and Prevention. Prevalence and impact of arthritis by race and ethnicity: United States, 1989–1991. MMWR Morb Mortal Wkly Rep 1996; 45: 3738.
  • 11
    Rasooly I, Papageorgiou AC, Badley EM. Comparison of clinical and self-reported diagnosis for rheumatology outpatients. Ann Rheum Dis 1995; 54: 8502.
  • 12
    Centers for Disease Control and Prevention. Prevalence and impact of chronic joint symptoms: seven states, 1996. MMWR Morb Mortal Wkly Rep 1998; 47: 34551.
  • 13
    Centers for Disease Control and Prevention. Health-related quality of life among adults with arthritis: behavioral risk factor surveillance system. 11 states, 1996–1998. MMWR Morb Mortal Wkly Rep 2000; 49: 3669.
  • 14
    Center for Disease Control and Prevention. Prevalence of arthritis: United States, 1997. MMWR Morb Mortal Wkly Rep 2001; 50: 3346.
  • 15
    Center for Disease Control and Prevention. Prevalence of self-reported arthritis or chronic joint symptoms among adults: United States, 2001. MMWR Morb Mortal Wkly Rep 2002; 51: 94850.
  • 16
    Center for Disease Control and Prevention. Prevalence of doctor-diagnosed arthritis and possible arthritis: 30 states, 2002. MMWR Morb Mortal Wkly Rep 2004; 53: 3836.
  • 17
    Bombard JM, Powell KE, Martin LM, Helmick CG, Wilson WH. Validity and reliability of self-reported arthritis Georgia senior centers, 2000–2001. Am J Prev Med 2005; 28: 2518.
  • 18
    Sacks JJ, Harrold LR, Helmick CG, Gurwitz JH, Emani S, Yood RA. Validation of a surveillance case definition for arthritis. J Rheumatol 2005; 32: 3407.
  • 19
    Riihimaki H. Musculoskeletal disorders. In: Ahrens W, Pigeot I, editors. Handbook of epidemiology. Berlin: Springer; 2005.
  • 20
    South Australian Department of Human Services. The Health Omnibus Survey (HOS) methodology. Adelaide (Australia): SADHS; 2002.
  • 21
    Australian Bureau of Statistics. ASCO: Australian Standard Classification of Occupations. 2nd ed. Canberra (Australia): ABS; 2001.
  • 22
    Australian Bureau of Statistics. National Health Survey: summary of results. Canberra (Australia): ABS; 2001.
  • 23
    Ware JE, Kosinski M, Gandek B. SF-36 health survey: manual and interpretation guide. 2000 ed. Lincoln: Quality Metric; 1993.
  • 24
    Hawthorne G, Richardson J, Osborne R. The Assessment of Quality of Life (AQOL) Instrument: a psychometric measure of health-related quality of life. Qual Life Res 1999; 8: 20924.
  • 25
    Osborne RH, Hawthorne G, Lew EA, Gray LC. Quality of life assessment in the community-dwelling elderly: validation of the Assessment of Quality of Life (AQOL) Instrument and comparison with the SF-36. J Clin Epidemiol 2003; 56: 13847.
  • 26
    Whitfield K, Buchbinder R, Segal L, Osborne RH. Parsimonious and efficient assessment of health-related quality of life in osteoarthritis research: validation of the Assessment of Quality of Life (AQOL) instrument. Health Qual Life Outcomes 2006; 4: 19.
  • 27
    Statistical package for the social sciences: version 15.0. Chicago: SPSS; 2006.
  • 28
    Tabachnick BG, Fidell LS. Using multivariate statistics. 2nd ed. Needham Heights (MA): Allyn & Bacon; 2001.
  • 29
    Hawthorne G, Osborne R. Population norms and meaningful differences for the Assessment of Quality of Life (AQOL) measure. Aust N Z J Public Health 2005; 29: 13642.
  • 30
    Feinglass J, Lee C, Durazo-Arvizu R, Chang RW. Health status, arthritis risk factors, and medical care use among respondents with joint symptoms or physician diagnosed arthritis: findings from the 2001 Behavioral Risk Factor Surveillance System. J Rheumatol 2005; 32: 1306.
  • 31
    Charles ST, Gatz M, Pedersen NL, Dahlberg L. Genetic and behavioral risk factors for self-reported joint pain among a population-based sample of Swedish twins. Health Psychol 1999; 18: 64454.
  • 32
    Huang C, Ross PD, Lydick E, Wasnich RD. Factors associated with joint pain among postmenopausal women. Int J Obes Relat Metab Disord 1997; 21: 34954.
  • 33
    Adamson J, Ebrahim S, Dieppe P, Hunt K. Prevalence and risk factors for joint pain among men and women in the West of Scotland Twenty-07 study. Ann Rheum Dis 2006; 65: 5204.
  • 34
    Feinglass J, Neilson C, Lawther T, Rowland WC. Chronic joint symptoms and prior arthritis diagnosis in community surveys: implications for arthritis prevalence estimates. Public Health Rep 2003; 118: 2309.
  • 35
    Busija L, Hollingsworth B, Buchbinder R, Osborne RH. Role of age, sex, and obesity in the higher prevalence of arthritis among lower socioeconomic groups: a population-based survey. Arthritis Rheum 2007; 57: 55361.
  • 36
    Schouten JS, van den Ouweland F, Valkenberg HA. A 12 year follow up study in the general population on prognostic factors of cartilage loss in osteoarthritis of the knee. Ann Rheum Dis 1992; 51: 9327.
  • 37
    Felson DT, Anderson JJ, Naimark A, Walker AM, Meenan RF. Obesity and knee osteoarthritis: the Framingham study. Ann Intern Med 1988; 109: 1824.
  • 38
    Abell JE, Hootman JM, Helmick CG. Prevalence and impact of arthritis among nursing home residents. Ann Rheum Dis 2005; 63: 5914.
  • 39
    Alonso J, Ferrer M, Gandek B, Ware JE Jr, Aaronson NK, Mosconi P, et al. Health-related quality of life associated with chronic conditions in eight countries: results from the International Quality of Life Assessment (IQOLA) Project. Qual Life Res 2004; 13: 28398.
  • 40
    Ayis S, Ebrahim S, Williams S, Juni P, Dieppe P. Determinants of reduced walking speed in people with musculoskeletal pain. J Rheumatol 2007; 34: 190512.
  • 41
    Blay SL, Andreoli SB, Gastal FL. Chronic painful physical conditions, disturbed sleep and psychiatric morbidity: results from an elderly survey. Ann Clin Psychiatry 2007; 19: 16974.
  • 42
    Verhaak PF, Kerssens JJ, Dekker J, Sorbi MJ, Bensing JM. Prevalence of chronic benign pain disorder among adults: a review of the literature. Pain 1998; 77: 2319.
  • 43
    Theis KA, Murphy L, Hootman JM, Helmick CG, Yelin E. Prevalence and correlates of arthritis-attributable work limitation in the US population among persons ages 18–64: 2002 National Health Interview Survey Data. Arthritis Rheum 2007; 57: 35563.
  • 44
    Dunlop DD, Manheim LM, Song J, Chang RW. Arthritis prevalence and activity limitations in older adults. Arthritis Rheum 2001; 44: 21221.
  • 45
    Bish CL, Blanck HM, Maynard LM, Serdula MK, Thompson NJ, Khan LK. Activity/participation limitation and weight loss among overweight and obese US adults: 1999 to 2002 NHANES. MedGenMed 2007; 9: 63.
  • 46
    Zoico E, Di Francesco V, Guralnik JM, Mazzali G, Bortolani A, Guariento S, et al. Physical disability and muscular strength in relation to obesity and different body composition indexes in a sample of healthy elderly women. Int J Obes Relat Metab Disord 2004; 28: 23441.
  • 47
    Christensen R, Bartels EM, Astrup A, Bliddal H. Effect of weight reduction in obese patients diagnosed with knee osteoarthritis: a systematic review and meta-analysis. Ann Rheum Dis 2007; 66: 4339.
  • 48
    Snibbe JC, Gambardella RA. Treatment options for osteoarthritis. Orthopedics 2005; 28 Suppl 1: 21520.
  • 49
    Coleman S, Conroy H, Rowley E, Fabling M, Briffa K, Carroll G, et al. Long term effects of education in the primary health care setting using behaviour modification, and specific exercises in patients with osteoarthritis of the knee: reduced pain, improved mobility and improved quality of life [abstract]. Arthritis Rheum 2002; 46 Suppl: S1989.
  • 50
    Buchbinder R, Goel V, Bombardier C. Lack of concordance between the ICD-9 classification of soft tissue disorders of the neck and upper extremity and chart review diagnosis: one steel mill's experience. Am J Ind Med 1996; 29: 17182.