Article Text


Benchmarking: the five year outcome of rheumatoid arthritis assessed using a pain score, the Health Assessment Questionnaire, and the Short Form-36 (SF-36) in a community and a clinic based sample


BACKGROUND Treatment, and therefore outcome, of rheumatoid arthritis (RA) will improve in the next few years. However, improvement in outcome can only be judged against the probability of certain outcomes with current conventional treatment.

AIM To document the five year outcome of RA in the late 1990s.

SETTING Norfolk Arthritis Register (NOAR).

DESIGN Longitudinal observational cohort study.

METHODS 318 patients with recent onset inflammatory polyarthritis recruited by NOAR in 1990–91 completed five years of follow up. Four groups were assessed: the whole cohort, all those referred to hospital, those who satisfied criteria for RA at baseline, and those referred to hospital who satisfied criteria for RA at baseline. Outcome was assessed with a visual analogue scale for pain, the Health Assessment Questionnaire (HAQ), and the Short Form-36 (SF-36).

RESULTS Of the RA hospital attenders, 50% had a visual analogue scale pain score of 5 cm or less and an HAQ score of 1.125 or less. SF-36 scores were reduced in all domains. Results are presented as cumulative percentages.

CONCLUSIONS These results can be used for comparison and to set targets for improvement.

  • rheumatoid arthritis
  • outcome
  • Health Assessment Questionnaire
  • Short Form-36

Statistics from

The treatment of rheumatoid arthritis (RA) is passing through exciting times. The past 12 months has seen the launch in the UK of a new class of non-steroidal anti-inflammatory drugs (COX-2 specific inhibitors),1 a new second line agent (leflunomide),2 3 and a new category of drugs—biological agents designed to block the action of key cytokines.4 5These drugs, used in conjunction with other established treatments, targeted at appropriate patients early in disease, should lead to a measurably improved outcome for people with RA. However, in order to establish in the future that outcome has improved, it is important to know “where we started from”—that is, the probability of certain outcomes with current conventional treatment. This paper describes the five year outcome of RA, measured using a pain score and two validated and widely used self administered questionnaires on physical function and quality of life, in a community based sample recruited in 1990–91. The results are presented for the whole sample and for the subgroup who were referred to hospital.


The Norfolk Arthritis Register (NOAR) aims at recruiting all adults (aged 16 and over) who consult a general practitioner (GP) in the former Norwich Health Authority with swelling of two or more joints lasting for at least four weeks and with an onset since l January 1989. The protocol has been described in detail elsewhere.6 In summary, referred patients are assessed by a metrologist, usually within two weeks of notification. The metrologist conducts a structured interview, examines the joints and takes blood for rheumatoid factor estimation. Patients are reviewed annually. The 1987 ACR criteria for RA7 are applied at baseline and at each annual assessment. They are applied “cumulatively”—that is, if a patient ever satisfies a particular criterion that result is carried forward to subsequent assessments.

This “benchmarking” exercise focused on three important aspects of disease outcome at five years from registration: pain, physical function, and health related quality of life (HRQoL). Pain was measured with a 10 cm visual analogue scale (VAS). The pain score was recorded to the nearest centimetre (range 0–10). Physical function was measured by the Stanford Health Assessment Questionnaire (HAQ),8modified for use by British patients.9 The HAQ measures functional ability in eight domains. It gives a score ranging from zero (no disability) to three (severe disability). HRQoL was assessed by the UK version of the Short Form-36 (SF-36).10 11 The SF-36 comprises 36 questions covering eight health domains: physical function, role limitations due to physical problems, bodily pain, general health, vitality, social functioning, role limitations due to emotional problems, and mental health. The score for each domain is transformed to a scale ranging from zero (poor health) to 100 (good health). As recommended by the developers of the SF-36,12missing values were imputed from the average of the completed items if over half the items for that domain had been completed.

Outcome measures completed by patients were chosen in order to avoid the issue of interobserver variation in measures such as joint counts, and so to facilitate comparison with other patient cohorts.


The study group was derived from 482 patients notified to the NOAR in 1990 and 1991. Forty nine were subsequently excluded because they were given a diagnosis other than RA, psoriatic arthritis, viral arthritis, or undifferentiated inflammatory polyarthritis by a rheumatologist. The remaining 433 patients were followed up for five years, during which time 44 (10.2%) died, 47 (10.9%) declined further follow up, and 24 (5.5%) were lost to follow up. Thus 318/433 (73.4%) patients completed five years of follow up. Of these, 302 (95.0%) completed a VAS for pain, 317 (99.7%) completed an HAQ, and 308 (96.9%) completed all or part of the SF-36 at the fifth anniversary assessment. Fifty one (16.6%) SF-36 forms were incomplete, but in 20 cases the missing data could be imputed. Two hundred and seventy seven patients (87.1%) provided sufficient information to calculate a score for each of the eight domains of the SF-36.

Of the 318 patients who completed five years follow up, 238 (74.8%) had been referred to hospital during the follow up period. Just over half of the group satisfied the 1987 American College of Rheumatology (ACR) criteria for RA7 at the time of notification to NOAR and 237 (74.5%) satisfied the ACR criteria when applied cumulatively. One hundred and thirty eight (86.3%) of the 160 patients who satisfied criteria for RA at baseline were referred to hospital. During the follow up period, 169/318 (53.1%) patients received second line treatment or steroids, or both. We looked at the outcome at five years for four groups: the whole cohort (n=318); all those referred to hospital (n=238); those who satisfied the ACR criteria for RA at baseline (n=160); and patients with RA referred to hospital (n=138).


Data were analysed with the Statistical Package for the Social Sciences (SPSS)13 and Stata.14 As the distribution of the scores for three domains (role-physical, role-emotion, and mental health) were highly skewed, median values (and interquartile range (IQR)) are reported for all eight domains. NOAR data were compared with data from a large UK population sample (16 054 adults), the 1996 Health Survey for England.15


Table 1 shows the baseline characteristics of the 433 patients with inflammatory polyarthritis who were recruited in 1990–91. The median age at onset was 56 years (IQR 42–68) and two thirds of the cohort were women.

Table 1

Patient characteristics at baseline (n=433)

Compared with those who completed five years follow up, those who died were older at the time of symptom onset (median age 70.0 yearsv 54.5 years) and more likely to be male (54.5% v 33.0%) (table 2). Patients who subsequently died also had a higher median HAQ (1.375; IQR 0.375–2.06) at registration than those who were followed up for five years (0.75; IQR 0.25–1.50. Although those who died were more likely to be seropositive, they were not more likely to satisfy the ACR criteria for RA at baseline. Those who withdrew from the study or were lost to follow up did not differ in age or sex from those who completed the follow up period; they did have milder disease as measured by the number of swollen joints and the proportion who satisfied the ACR criteria at the time of notification to NOAR (table 2).

Table 2

Comparison of the baseline characteristics of study completers with study non-completers

The four groups analysed did not differ in their reported level of pain at five years (table 3). Figure 1 shows the cumulative proportion of each of the four patient groups who had a VAS pain score of a particular value or below. Thus, for example, 50% of hospital attenders with RA had a VAS pain score of 5 cm or more and 50% of all the other groups had a VAS pain score of 4 cm or more.

Table 3

Characteristics of study completers at five years

Figure 1

Visual analogue scale (VAS) pain score at five years.  

There was, however, a difference in HAQ scores between the groups (table 3). Figure 2 shows the cumulative proportion of each of the four patient groups who had an HAQ score of a particular value or below. Thus, for example, 30% of hospital attenders with RA had an HAQ score of 0.50 or less. Fifty per cent of the patients with inflammatory polyarthritis had an HAQ score of 0.75 or more at five years. In contrast, in the hospital attenders with RA, 50% had an HAQ of 1.125 or more. It is difficult to interpret precise values of the HAQ. However as a guide, an HAQ score of one or more is often used to represent “moderate” disability.16 17 Further details of the functional outcome and predictors of disability in this cohort have been published elsewhere.16 17

Figure 2

Health Assessment Questionnaire (HAQ) scores at five years. IP = inflammatory polyarthritis.

All groups showed some impairment in HRQoL as measured by the SF-36 (table 4). The most seriously affected domain was the role-physical domain. The only difference between the groups was in the role-emotional domain. In the RA hospital attenders subgroup and the RA subgroup, 50% of patients had a score of zero in the role-physical domain. Although the score for each of the SF-36 domains ranges from 0 (poor health) to 100 (good health), the number of steps between the top and bottom of this range varies. For example, in the role-physical domain, the scoring system is such that it is only possible for the patient to score 0, 25, 50, 75, or 100. However, for other domains (for example, physical function and vitality) there are more than 20 possible scores. Figure 3 shows the results of the NOAR patients compared with those from a large sample of the UK population (controlled for age and sex).15 Compared with the general population, the NOAR patients (except for the RA hospital attenders) had little impairment in the mental health and role-emotional domains.

Table 4

SF-36 domain scores at the fifth anniversary assessment

Figure 3

Comparison of NOAR SF-36 domain scores with UK population normative data. IP = inflammatory polyarthritis.


Theoretically the patients presented in this report should include all new cases of inflammatory polyarthritis (and RA) who presented to primary care in the Norwich Health Authority in 1990–91. There will of course be underascertainment—mainly due to failure by GPs to notify the NOAR of patients who were not subsequently referred to hospital. We do not know the degree of underascertainment, though we are currently conducting a population based prevalence survey of RA in this area which should provide some insight into the number of missing cases. Some cases with an onset in 1990–91 will have presented to primary care after 31 December 1991 and so this cohort is an inception cohort of early inflammatory polyarthritis that presented in 1990–91.

There have been relatively few reports of the five year outcome of RA using the HAQ18 19 20 and none using either a VAS pain score or the SF-36. Kvien et al used a VAS pain score in a survey of 1552 subjects on a community register of all hospital diagnosed cases of RA in Oslo.21 The mean VAS in the 1030 respondents (who had a mean disease duration of 13 years) was 4.6 cm (SD 0.08). This is similar to the median VAS of 5 cm found in our study, suggesting that pain score may vary little with disease duration. In another inception cohort study from Norway the mean HAQ at five years was 0.91 (SD 0.65) and 40% of patients had an HAQ score >1.20 The median HAQ score at five years in patients recruited at each of the nine UK centres contributing to the Early Rheumatoid Arthritis Study (ERAS) ranged from 0.8 to 1.3.19 In seven of the nine centres the median HAQ was either 1.0 or 1.1. The median HAQ for the NOAR RA hospital attenders was 1.125.

Although there have been a number of publications (mainly clinical trials) which have used the SF-36 in patients with RA, few have reported the actual scores for each domain. Two of these studies involved UK hospital attenders (table 5). They represent the whole spectrum of disease duration, with a substantially higher median duration than the NOAR patients. This is reflected in the poorer median scores for physical function and bodily pain compared with the NOAR RA hospital scores. The patients on the Oslo community register of patients with hospital diagnosed RA, who also had a greater disease duration than the NOAR community patients with RA, generally had poorer scores. A population survey in Australia which asked respondents whether they had “ever been diagnosed as having RA by a doctor” found better scores in all domains of the SF-36 than any other published study of RA24 (table 6). This raises the possibility of diagnostic misclassification.

Table 5

Comparison of SF-36 data from NOAR with other studies of UK hospital patients with RA

Table 6

Comparison of SF-36 data from NOAR with other community based studies of RA

We believe that the rate of referral and the management of hospital referred cases was “typical” for the UK at the time (1990–91). During the follow up period 169 (53%) patients received second line drugs or steroids. Sulfasalazine was the drug of first choice, though methotrexate became more popular during the follow up period. These data may, therefore, be regarded as typical of the five year outcome of RA and inflammatory polyarthritis in the UK in the late 1990s. Figures1 and 2 could be used to set targets for better treatment in the future—and this in turn offers possibilities for RA to be included in National Programmes for Health Improvement. For example, from fig 2, it can be seen that 40% of hospital attenders with RA have an HAQ score <0.75 at five years. It would then be possible to set a target that 60% of such patients should have an HAQ score <0.75 at some future date. Similarly, this could provide a standard against which other departments could assess the outcome of their patients. Any differences in outcome between centres would then offer avenues for further research.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.