METHODS

Location

The Central Finland District is located in the southern part of Finland. Its population of 263 869 (in the year 2000) is 5% of the total population of Finland.

Patients with rheumatoid arthritis

Jyväskylä Central Hospital is the only rheumatology centre in the Central Finland District. All new patients with rheumatoid arthritis are referred to this centre for diagnostic studies and initiation of treatment. Most patients with severe rheumatoid arthritis visit the outpatient or inpatient clinic regularly. The Central Finland rheumatoid arthritis database includes demographic measures, treatments, and outcomes of all patients with rheumatoid arthritis seen in the clinic since January 1993. It is updated daily by a research nurse. By June 2000, the database contained 1763 patients, 1495 of whom were still alive. These 1495 patients were mailed a questionnaire. The questionnaire results were entered into the rheumatoid arthritis database and analysed along with the patient’s disease characteristics and treatments.

Controls

To obtain a population sample, the names and addresses of 2000 people who were at least 30 years old and living in the district were requested from the Ministry of Social Affairs and Health. The sample was drawn from the Population Registry database that includes a personal identification code, date of birth, demographic data, name, and address of all individuals living in Finland. The population sample was designed to have a mean age of 55 years and to include 70% women, so as to be similar to the rheumatoid population, but was otherwise random. The sampling was done by Statistics of Finland. This organisation operates administratively under the Ministry of Finance but is fully and independently responsible for its statistics. Statistics of Finland obtains the majority of the data from diverse administrative registers, and produces two thirds of all government statistics in Finland (www.stat.fi).

Study design

A questionnaire was mailed to the 1495 patients with rheumatoid arthritis and the 2000 controls in June 2000. A reminder was sent eight weeks later to non-responder controls.

Functional status in activities of daily living was assessed by the Finnish version of the HAQ,¹³ and included 20 questions in eight categories: dressing, rising, eating, walking, grooming, reaching, gripping, and performing errands. The response alternatives available were 0, no difficulty; 1, some difficulty; 2, much difficulty; 3, unable to do. The sum of the highest response in each category was divided by 8 to form a score from 0 to 3. Perceived pain and global status were assessed on a 100 mm visual analogue scale (VAS).

Information on the presence of comorbidities was sought, including hypertension, coronary artery disease, other heart disease, asthma, chronic bronchitis, chronic kidney disease, peptic ulcer, inflammatory bowel disease, diabetes, thyroid disease, cancer, epilepsy, stroke, Parkinson’s disease, psoriasis, chronic leg ulcers, ankylosing spondylitis, osteoarthritis, fibromyalgia, chronic back pain, musculoskeletal trauma, mental illness, and alcoholism. The number of comorbidities was calculated as a plain sum of these conditions.

Other data collected included date of birth, height and weight for body mass index (BMI; kg/m²), sex, length of education, and information on smoking status.

End point measurements

Two years after administration of the questionnaire, the vital status of the subjects was ascertained from a database in the Population Register Centre of Finland, using personal identification codes, which have been used in all population registers in Finland since 1967.

Statistical analysis

Data analyses were done using SPSS 11.0 (SPSS Inc, Chicago, Illinois, USA) and STATA 7.0 software (STATA Corporation, College Station, Texas, USA). The results for continuous variables are given as mean (SE), and for dichotomous variables as percentages. The differences among the groups were computed by χ² tests (for dichotomous data) or Student’s two tailed t test (for continuous data) for independent samples, as appropriate.

Univariate and multivariate Cox regression models were used to explore the association of risk factors with mortality, which were computed separately for patients with rheumatoid arthritis and for the community controls. In each case the final model included variables that were significant in univariate analysis, adjusted for age and sex (data not shown). Thus in patients with rheumatoid arthritis the final model included disease duration, rheumatoid factor status, HAQ, VAS for pain and global status, number of comorbidities, age, and sex as covariates. In the community controls the final model included VAS for pain and global status, HAQ, age, and sex. The Cox regression model was used to compare the mortality in patients with rheumatoid arthritis with community controls, adjusted for age, sex, and HAQ score.

Kaplan–Meier survival curves were used to illustrate mortality according to the response to the HAQ in four groups: patients with rheumatoid arthritis with a HAQ score of ⩾1; patients with rheumatoid arthritis with a HAQ score of <1; community controls with a HAQ score of ⩾1; and community controls with a HAQ score of <1. A cut off point of 1 (on the scale of 0–3 of the HAQ) was chosen as it is clinically relevant and indicates at least some difficulty in most of the activities of daily living.

The study was approved by the ethics committee of Jyväskylä Central Hospital, and Population Register Centre of Finland.

RESULTS

Patients

In all, 1095 subjects with rheumatoid arthritis (73% of 1495) returned a completed questionnaire. The mean age of the respondents was 62.4 years (range 19 to 96), 71% were female, mean disease duration was 11.3 years, 69% (of 1013 patients tested) were positive for rheumatoid factor, and 60% had erosions on radiography of their hands or feet (table 1).

Controls

In all, 1530 control subjects returned the questionnaire (77% of 2000), 1490 of whom completed the HAQ. The mean age of the respondents was 55.5 years (range 30 to 91), and 72% were female (table 1).

Mortality

Over the two year period, the number of deaths was 41 (10.1%) among 404 patients with rheumatoid arthritis with a baseline HAQ score of ⩾1, and 16 (2.3%) among 691 patients with a HAQ score of <1 (p<0.001). The corresponding figures for the community controls were 20 (13.6%) of 147, and 14 (1.0%) of 1343 (p<0.001) (table 1).

In a Cox regression model for the patients with rheumatoid arthritis, a higher HAQ score was an independent predictor of mortality, with a hazard ratio of 2.73 (95% confidence interval, 1.86 to 4.02) (p<0.001). Other independent predictors of mortality were male sex (p = 0.001), older age (p = 0.004), number of comorbidities (p = 0.010), and positive rheumatoid factor (p = 0.017) (table 2). In the community controls, a higher HAQ score was an independent predictor of mortality, with a hazard ratio of 2.75 (1.61 to 4.70) (p<0.001) (table 3).

Patients with rheumatoid arthritis were more likely to die over two years than the community controls in an unadjusted analysis (table 1). In a Cox regression model adjusted for age and sex, mortality in patients with rheumatoid arthritis was also higher than in the community controls (p = 0.027, data not shown). However, when the functional capacity of the subjects was included in the model, having rheumatoid arthritis was not an independent predictor of mortality (table 4), which indicates that similar levels of disability are associated with similar likelihood of mortality, regardless of having rheumatoid arthritis.

Figure 1 shows the Kaplan–Meier estimates of the probability of death in patients with rheumatoid arthritis and community controls, according to the baseline HAQ score. Subjects with higher HAQ scores of ⩾1 had the worst prognosis in both of the groups.

DISCUSSION

HAQ scores have been shown to predict severe outcomes of rheumatoid arthritis including mortality, work disability, and joint replacement surgery.^5–7 Our study confirms previous reports that the HAQ predicts mortality in patients with rheumatoid arthritis. However, the novel finding of our study is that the HAQ also predicts mortality in the general population. In the Cox regression model, the HAQ appeared to be an independent predictor of death in the general population, with a hazard ratio of 2.75 (1.61 to 4.70), similar to that of patients with rheumatoid arthritis, in whom the hazard ratio was 2.73 (1.86 to 4.02). In other words, the risk of death increased 2.7 times for each one point increase in the HAQ score (on a scale of 0 to 3) in both groups. Similar levels of disability were associated with a similar likelihood of mortality over two years, regardless of the diagnosis of rheumatoid arthritis (table 4).

The HAQ includes questions concerning the subject’s experience in coping with simple activities of daily living. In patients with rheumatoid arthritis, the HAQ has been shown to be associated with pain and psychological variables, and only weakly with radiographic erosion scores or joint deformities.^14,15 Overall, pain and functional loss are weakly associated with laboratory tests and radiographs in any of the chronic musculoskeletal diseases such as osteoarthritis¹⁶ and chronic low back pain.^17,18 It therefore appears that a self report questionnaire concerning daily activities captures information that is not available from any other source but the patient.

Healthy habits were associated with slow progression of disability and better survival in elderly runners v community controls,¹⁹ and in university alumni.²⁰ In these studies, the progression of disability was measured by the HAQ. The results of the present study are in line with previous findings, and extend the use of the HAQ to apply all people over 30 years of age in the community, while in previous studies the HAQ was applied only in elderly people.

A limitation of our study is that we did not explore the reasons for the baseline disability in depth. No physical examination was undertaken, and no high technology data were obtained. One might also argue that the sickest people die, and the results of the present study do not provide any novel information. Nevertheless, the study shows that a simple and inexpensive one page self report questionnaire regarding activities of daily living, which only takes between one and five minutes to complete, is a feasible tool for predicting mortality at a population level over a two year period, not only in patients with rheumatoid arthritis but also in the general population. The likelihood of dying appears to be similar in people with similar levels of disability according to the HAQ, regardless of whether or not they have rheumatoid arthritis.