Changes in Health Assessment Questionnaire disability scores over five years in patients with rheumatoid arthritis compared with the general population

Authors


Abstract

Objective

To analyze longitudinal data over 5 years for changes in Health Assessment Questionnaire (HAQ) scores in patients with rheumatoid arthritis (RA) and age- and sex-matched controls from the general population.

Methods

In 2000 and 2005, identical self-report questionnaires were mailed to a cohort of patients with RA and control cohort from the community. The questionnaire included the HAQ, which was used to assess functional status. Changes in HAQ scores over 5 years were analyzed.

Results

In 2000, 73% of 1,495 patients with RA and 77% of 2,000 general population controls responded to the questionnaire. In 2005, 84% of 2,022 patients with RA and 77% of 1,817 controls responded. A total of 863 patients with RA and 1,176 community controls responded in both 2000 and 2005 and were included in the analyses. Mean baseline HAQ scores were significantly higher in patients with RA than in controls (0.71 versus 0.17; P < 0.001). Over 5 years, the HAQ scores increased by 0.01 units per year in both the RA cohort and the community population; in both cohorts, the net change was primarily attributable to individuals over age 70 years. Changes in HAQ scores were similar in patients and controls who had low HAQ scores at baseline. Female patients with baseline HAQ scores of ≥0.5 had less potential for improvement than did controls. Among subjects in both groups who had HAQ scores >2, death was a common outcome over the next 5 years.

Conclusion

Currently, progression of functional disability among patients with RA and among persons in the general population is largely explained by the aging process. Our results showing stable function scores over 5 years in most patients with RA who are younger than age 70 years provide further evidence of improved status of RA patients today compared with the major declines observed in previous decades.

Functional status is the most significant prognostic measure of long-term outcomes in rheumatoid arthritis (RA), such as work disability (1–3), mortality (4–6), costs (7), the need for joint replacement surgery (8), and loss of function (4, 9). Furthermore, patients with RA who have significant functional disability have a 3-fold increased risk of mortality compared with that in the general population; this risk is comparable with that in individuals in the general population in the highest quintile for systolic and diastolic blood pressure, cholesterol level, or pack-years of smoking (10). In addition, improvements in disease activity and Health Assessment Questionnaire (HAQ) (11) scores in patients with early RA are associated with favorable long-term outcomes (12), analogous to improved long-term outcomes achieved by normalization of blood pressure or serum cholesterol levels.

The HAQ was developed in the late 1970s to measure outcomes in patients with arthritis. Reference values in a normal population were not included in the initial analysis nor in almost all subsequent analyses. Such information would appear to be of considerable interest in view of evidence that poor physical function is a predictor of mortality in nonrheumatic conditions, including congestive heart failure (13) and acquired immunodeficiency syndrome (14, 15). In 2004, we reported results of a cross-sectional analysis of our population-based cohort and provided reference values for the HAQ (16), including evidence that poor physical function according to elevated HAQ scores is a predictor of early mortality in the general population, as in patients with RA (6).

In longitudinal studies of patients with RA who were monitored in the 1980s and 1990s, HAQ scores were estimated to progress by 0.01–0.03 units per year (17–19). However, as had been the case for cross-sectional data, longitudinal data concerning the change in HAQ scores in the general population compared with patients with RA have not been reported. The availability in Central Finland of longitudinal data for patients with RA and age- and sex-matched community-based controls provides an unusual opportunity to analyze changes in HAQ scores over 5 years.

PATIENTS AND METHODS

Location.

In 2004, the Central Finland Health Care District included a population of 267,182 individuals, representing 5% of the total population of Finland. The current study was approved by the Ethics Committee of Jyväskylä Central Hospital and by the Population Register Centre of Finland.

Patients with RA.

Jyväskylä Central Hospital is the only rheumatology center in the Central Finland Health Care District. All new patients with suspected RA are referred to this center for diagnostic studies and initiation of therapy. Most patients with persistent RA continue to receive care at this hospital, with regular visits to outpatient and daycare clinics. The Central Finland RA database includes demographic and clinical data on all patients with RA who have been seen in Jyväskylä Central Hospital since 1993. Since 1998, all patients with RA have been monitored through questionnaires that are mailed annually.

Community control subjects.

A control cohort of 2,000 individuals from the general population, matched for age and sex with patients in the RA cohort, was established in 2000. This sample was drawn from the Finnish Population Registry, with consent from the Ministry of Social Affairs and Health. In order to match the RA patient group by age and sex, the population sample was designed to have a mean age of 55 years and to include 70% women, but was otherwise random. All subjects in the control group were living in the Central Finland District in 2000.

Study design.

A self-report questionnaire was mailed to patients with RA and control subjects in June 2000. Functional status in activities of daily living was measured by the Finnish version of the HAQ (20). Pain and global health were assessed on a 100-mm visual analog scale. The presence of comorbidities, the frequency of physical exercise, as well as each subject's education level, height, weight, and smoking status were also queried.

Previous analyses of these data indicated that RA is associated with a >7-fold risk of disability compared with that in a general population (21). Patients with RA and control subjects with similar levels of functional disability according to the HAQ appeared to have a comparable likelihood of higher mortality over 2 years than did patients and controls with no functional disability (6). A history of smoking was shown to be associated with an increased risk of rheumatoid factor–positive RA in men (22). Furthermore, nonresponse to the questionnaire was found to be associated with increased mortality (23).

In June 2005, the questionnaire was mailed to all of the patients with RA who were still alive, and to the same community controls as in 2000, except those who had moved from the district or had died. The 2005 questionnaire was identical to the 2000 questionnaire.

Statistical analysis.

Results for continuous variables are presented as the mean ± SD. Generalized linear models were used to analyze differences in the change in HAQ scores over 5 years between the cohorts and age groups (<40, 40–49, 50–59, 60–69, 70–79, and 80+ years), and to determine whether differences exist in cohort-by-age interaction. The 95% confidence intervals were obtained using the bias-corrected bootstrapping procedure with 5,000 replications. Analyses were also performed with adjustment for baseline HAQ scores. In separate analyses, baseline HAQ scores were categorized arbitrarily as 0, 0.01–0.49, 0.50–0.99, 1.00–1.99, and ≥2.00.

RESULTS

Patients and control subjects.

In 2000, 1,095 (73%) of 1,495 patients with RA and 1,530 (77%) of 2,000 population controls responded to the questionnaire. In 2005, 1,705 (84%) of 2,022 patients with RA and 1,400 (77%) of 1,817 controls responded to the questionnaire (Table 1). The analyses of the change in HAQ scores over 5 years included 863 patients with RA and 1,176 community controls who responded to the questionnaire in both 2000 and 2005 and for whom all items of the HAQ were complete.

Table 1. HAQ scores in patients with RA and control subjects, according to their status in 2000 and 2005*
GroupHAQ score
20002005
  • *

    Values are the mean ± SD. HAQ = Health Assessment Questionnaire; RA = rheumatoid arthritis.

  • Twenty-eight patients in the population sample in 2000 were also included in the RA group and were analyzed as part of the RA group in 2005.

All RA patients contacted in 2000 (n = 1,495)  
 Responders in 2000 (n = 1,095)  
  Responders in 2005 (n = 863)0.71 ± 0.710.76 ± 0.78
  Nonresponders in 2005 (n = 93)0.96 ± 0.88
  Dead in 2005 (n = 139)1.48 ± 0.98
 Nonresponders in 2000 (n = 400)  
  Responders in 2005 (n = 188)0.90 ± 0.84
  Nonresponders in 2005 (n = 133)
  Dead in 2005 (n = 72)
  Missing in 2005 (n = 7)
RA patients included after 2000 (n = 745)  
 Long-term RA  
  Responders in 2005 (n = 177)0.82 ± 0.76
  Nonresponders in 2005 (n = 26)
 Newly diagnosed RA  
  Responders in 2005 (n = 477)0.49 ± 0.58
  Nonresponders in 2005 (n = 65)
Population controls (n = 2,000)  
 Responders in 2000 (n = 1,530)  
  Responders in 2005 (n = 1,209)0.17 ± 0.410.22 ± 0.49
  Nonresponders in 2005 (n = 192)0.32 ± 0.64
  Dead in 2005 (n = 91)1.38 ± 1.15
  Missing in 2005 (n = 10)0.23 ± 0.51
  Population sample and RA group (n = 28)0.72 ± 0.70
 Nonresponders in 2000 (n = 470)  
  Responders in 2005 (n = 191)0.27 ± 0.55
  Nonresponders in 2005 (n = 225)
  Dead in 2005 (n = 36)
  Missing in 2005 (n = 18)

Mean HAQ scores.

Table 1 shows the mean ± SD HAQ scores of all patients with RA and control subjects who were included in the study in 2000 and/or 2005. Mean HAQ scores for functional disability were statistically significantly higher in patients with RA compared with controls, including individuals who responded both in 2000 (0.71 versus 0.17) and 2005 (0.76 versus 0.22), those who responded in 2000 but did not respond in 2005 (0.96 versus 0.32), and those who did not respond in 2000 but responded in 2005 (0.90 versus 0.27) (all P < 0.001, after adjusting for age and sex). Patients with RA and population-based controls who responded in 2000 and subsequently died had similar baseline HAQ scores (mean 1.48 versus 1.38; P = 0.29), which were higher than the scores for other subgroups in these cohorts.

Changes in HAQ scores in the cohorts over 5 years.

Changes in HAQ scores over 5 years were analyzed in 863 patients with RA and 1,176 subjects in the general population (Table 1). The overall increase in the mean HAQ score was 0.01 units per year in patients with RA as well as in controls. Mean HAQ scores in the cohort of patients with RA increased very slightly over 5 years (from 0.71 to 0.76), as did those in the control population (from 0.17 to 0.22) (Table 1). Almost all of the mean change was attributable to persons older than age 70 years, in both the RA and control cohorts (Figures 1 and 2).

Figure 1.

Change in Health Assessment Questionnaire (HAQ) scores over 5 years in female patients with rheumatoid arthritis (RA) and control subjects, according to age. Values are the mean and 95% confidence interval. Broken line represents baseline.

Figure 2.

Change in Health Assessment Questionnaire (HAQ) scores over 5 years in male patients with rheumatoid arthritis (RA) and control subjects, according to age. Values are the mean and 95% confidence interval.

The potential for change in HAQ scores over 5 years was greater in elderly women with RA compared with control subjects (Figure 1). Similar results were seen in men, but the confidence intervals were wider due to the lower number of individuals (Figure 2). In women, the mean change in HAQ scores was similar between patients with RA and controls (P = 0.64), while statistically significant differences were seen according to age (P < 0.001) and cohort-by-age interaction (P = 0.03). In men, the cohort effect (P = 0.80) and cohort-by-age interaction (P = 0.12) were not significant, in contrast to a significant age effect (P = 0.004). After adjustment to the baseline HAQ score, the cohort effect was significant in women (P < 0.001) but not in men (P = 0.17) (Figure 3).

Figure 3.

Change in Health Assessment Questionnaire (HAQ) scores over 5 years in patients with rheumatoid arthritis (RA) and control subjects, according to the baseline HAQ scores, adjusted for age. Values are the mean and 95% confidence interval. Broken line represents baseline.

Changes in HAQ scores in individuals over 5 years.

Over 5 years, HAQ scores among female patients with RA improved in 27%, deteriorated in 28%, and remained the same in 45%, compared with improvement in 10% of female controls, deterioration in 15%, and stability in 75%. In male patients with RA the corresponding values were 18%, 29%, and 53%, and in male controls the corresponding values were 8%, 13%, and 79%.

Effect of baseline HAQ scores.

The age-adjusted change in HAQ scores was similar in patients and controls who had low baseline HAQ scores. Among individuals with higher baseline HAQ scores, patients with RA had less improvement than did controls; the difference was statistically significant in women with baseline HAQ scores ≥0.5 (Figure 3). Death was a common outcome among individuals who had baseline HAQ scores >2; 40% of patients with RA and 62% of controls with HAQ scores of >2 at baseline died over 5 years.

DISCUSSION

This study is the first to compare changes in HAQ scores over 5 years in patients with RA and age- and sex-matched controls from the general population living in the same geographic area. A primary observation is that the average HAQ disability score did not increase in groups of patients with RA or in the general population prior to age 70 years.

At baseline, patients with RA had statistically significantly higher levels of disability compared with controls. Although the change in HAQ scores was similar in RA patients and controls, the absolute level of disability at followup was higher in patients with RA, as expected.

Among patients with RA of long duration who were followed up during the 1980s, HAQ disability scores were estimated to increase by 0.02–0.03 units per year (24, 25). In a subset of the Arthritis, Rheumatism, and Aging Medical Information System (ARAMIS) database established in the 1970s, the average disability score increased by 0.01–0.02 units per year over a 10-year period (17). More recently (through the late 1990s), Wolfe and Choi (18, 19) reported an increase of 0.01–0.03 units per year among patients with RA living in Wichita, Kansas, and in the National Data Bank. In our data set, the overall increase in HAQ scores of 0.05 units over 5 years indicated an increase of 0.01 units per year, although almost all of the net change was attributable to individuals over age 70 years (both patients with RA and controls). It is possible that progression of disability is explained almost entirely by age in general at this time, because HAQ scores remained similar over 5 years in people younger than age 70 years. Whether the observed increase of 0.01–0.03 units in the HAQ score in previous studies reflects changes in the entire cohort or primarily in the oldest age groups is not known, because this topic was not examined in depth.

Lesser progression of disability in our RA cohort compared with that in previous studies of patients with RA may reflect lower baseline HAQ scores in our patients. In other studies, mean baseline HAQ scores were 1.4–2.0 (24, 25), 0.8–1.4 (17), and 1.0 (19), compared with 0.71 in the present study. The extent of disability has been shown to be predictive of future disability (4). In our study, patients with higher baseline HAQ scores showed less potential for improvement compared with controls (Figure 3). Furthermore, 40–62% of individuals with baseline HAQ scores >2 did not survive over the next 5 years, extending previous observations of the prognostic value of a poor HAQ score for substantially higher mortality (6).

Vita et al reported cumulative lifetime disability in a non-RA population of elderly university alumni who had health risks recorded during midlife and late adulthood (26). Progression of disability was seen primarily after age 70 years, with the lowest progression in persons with a healthy lifestyle. Individuals who subsequently died had the highest levels of self-reported disability at baseline. Similar results were obtained from a 13-year study of elderly runners (27). In both of those studies, a healthy lifestyle was associated with postponed disability and mortality.

The current study has several limitations, the first of which is the relatively short observation period of 5 years. It will be of interest to observe whether, in the future, the potential for declines in HAQ scores continues to be seen primarily in older age groups. Another limitation of our study is nonresponsiveness and missing individuals, most of whom moved away from the district, although the response rates of 73–84% in both 2000 and 2005 are quite high for a survey study. A third limitation is that the control population did not consist of healthy individuals but rather comprised a random sample of individuals from the general population (who had any health condition except RA). Furthermore, the study was conducted in a single geographic district with a homogeneous population, and generalizability of the results must be tested in other settings.

Our findings of stable levels of disability are in contrast to findings from studies performed 20–30 years ago, including longitudinal observational studies of patients with early RA over 20–40 years (31, 32), during which most patients with RA experienced severe functional declines (4, 28–30). More recent studies performed over 5–10 years in cohorts with early RA showed varying results, with considerable progression of disability in many patients (33, 34), as well as preserved (35) and improved status over 5 years (36) compared with that at the time of presentation. In the Norfolk Arthritis Register study, HAQ scores remained stable over 5 years in patients with early inflammatory polyarthritis (37).

Currently, aggressive use of disease-modifying antirheumatic drugs (even before the era of biologic agents) is associated with improved clinical status of patients with RA compared with that of patients in previous decades, according to disease activity (38, 39), functional capacity (35, 39–41), radiographic severity scores (39, 42, 43), and other clinical measures (40), including lower mortality among patients whose disease responded to methotrexate (44, 45). Our results showing stable HAQ disability scores over 5 years in patients with RA who are younger than age 70 years provide further evidence of the improved clinical status of patients with RA compared with the major declines observed in such patients in previous decades.

Ancillary