Materials and methods

Patients

This study included 255 patients with inflammatory arthropathies and inclusion occurred during the period February 2005 to May 2006 at Diakonhjemmet Hospital, Oslo, Norway. The 255 patients underwent orthopaedic surgical treatment and responded to the mail surveys at baseline and at least one point of follow-up. Patients with the diagnosis of osteoarthritis were not included and we also excluded patients who underwent revision surgery, surgery because of fractures or secondary infections. Each patient was only included once (first intervention during the inclusion period). The distribution of diagnoses was as follows: RA (64.2%), juvenile RA (3.9%), ankylosing spondylitis (5.1%), psoriatic arthritis (7.4%) and other arthritides (12.0%). The mean baseline age was 57.5 years (SD 13.1) and 76.7% were women. Patients with RA underwent more replacement surgery than patients with other arthritides (30.4% vs 17.8%). The distribution of surgical procedures in the upper and lower limb was similar in RA and the other arthritides.

Age and gender were similar in the groups of patients who underwent surgery in the upper and lower limb. The patients who underwent replacement surgery were older (61.4 vs 56.0 years, p = 0.006) than patients who underwent other surgical interventions, but the sex ratio was similar. All patients were seen by a physiotherapist as part of the routine and, if needed, they were offered physiotherapy after the surgical procedure.

Data collection

The PRO were recorded at hospital admission and then by mail surveys 3, 6, 9 and 12 months after the surgical procedure. The booklet of questionnaires included both generic and disease-specific health status measures (arthritis impact measurement scales 2 (AIMS2), the health assessment questionnaire (HAQ), the short form 36 (SF-36), the utility instrument EQ-5D and four visual analogue scales (VAS), which addressed patient global assessment of disease, fatigue, general pain and pain in the specific joint undergoing surgery.

Instruments

AIMS2 is a multidimensional disease-specific measure that has been translated into many languages including Norwegian.10 11 The first 57 items of AIMS2 are broken into 12 scales: mobility (five items); walking and bending (five items); hand and finger function (five items); arm function (five items); self-care tasks (four items); household tasks (four items); social activity (five items); support from family and friends (five items); arthritis pain (five items); work (four items); level of tension (five items) and mood (five items). The scales may be combined into a five-component model reflecting the physical dimension, affect, symptoms, social interaction and role. The score of each scale ranges from 0 to 10 (10 represents worst health). AIMS2 was scored according to the AIMS2 users’ guide issued by the Boston University Arthritis Centre.

The HAQ is a disease-specific measure for RA and asks questions about the patient’s ability to perform activities of daily living.12 The HAQ includes questions assessing difficulty over the past week in 20 specific functions, grouped into eight categories: dressing and grooming; arising; eating; walking; personal hygiene; reaching; gripping and other activities. The responses to “are you able to do…” were scored 0–3 (without any difficulty  =  0, unable to do  =  3). The total HAQ score is the mean of the scores for the eight categories. Scores were adjusted for the use of assistive devices.

The SF-36 is a generic health status measure and has been used in a variety of conditions including RA, musculoskeletal disorders and patients who have undergone replacement surgery.13 14 15 The eight multi-item scales are as follows: physical functioning (10 items); role limitations due to physical health (four items); bodily pain (two items); general health (five items); vitality/energy/fatigue (four items); social functioning (two items); role limitations due to emotional problems (three items) and mental health (five items). SF-36 scales were scored according to published scoring procedures, each expressed with values from 0 to 100 (0  =  poor health).

SF-6D is a utility score which is derived from the SF-36. The eight dimensions in SF-36 are reduced to six in SF-6D. The level of severity in each dimension is determined by responses to related items in SF-36. The six dimensions are: physical function; role limitations; social function; pain; mental health and vitality.16 The utility score has a range from 0 to 1, but the lowest achievable value with SF-6D for living patients is 0.29.17

The EQ-5D is a standardised utility instrument with a three-level five-dimensional format.18 19 The EQ-5D includes the following dimensions: mobility; self care; usual activities; pain/discomfort and anxiety/depression. Negative values can also be achieved in living patients.17

Surgical procedures

The different types of surgical procedures are listed in appendix 1 (published online only as a supplementary file). The surgical procedures were categorised into replacement surgery and non-replacement surgery, and also into surgical procedures performed in upper versus lower limbs. The primary large joint replacement arthroplastics are referred to as replacement surgery in this article. For analytical purposes total replacements in the small joints (metacarpophalangeal, proximal interphalangeal and metatarsophalangeal) were categorised together with other surgical procedures than replacement surgery in the larger joints in order to be able to analyse homogenous groups according to the postoperative follow-up and rehabilitation.

Statistical analyses

Changes within patient groups during follow-up from baseline to 12 months were examined by paired sample t test. The standardised response mean (SRM) reflects the magnitude of an improvement (or deterioration) and was computed as the change from baseline to the 12-month follow-up divided by the standard deviation of the change score. The magnitude of the SRM was interpreted in terms of the thresholds introduced by Cohen20 for effect sizes: “trivial” (<0.20), “small” (>0.20<0.50), “moderate” (>0.50<0.80) or “large” (>0.80).

The SRM is comparable across the different instruments, independent of the instrument scales.

An independent sample t test was used to compare longitudinal changes between groups (replacement vs non-replacement surgery and surgery in upper vs lower limb). The last observation carried forward (LOCF) was used to replace missing values during follow-up. p Values equal to or below 0.05 were considered to be statistically significant. Correction for the number of tests was not performed as this was an explorative study. SPSS 14.0 and SPSS 15.0 were used in the analyses.

Results

Surgical procedures in upper versus lower limbs were performed in 42.4% and 57.6% of the 255 patients. Replacement surgery was performed in 25.6% of the patients and 75.0% of these procedures were performed in the lower limbs.

Significant improvement was seen for most of the dimensions of health (fig 1), but the magnitude of improvement differed considerably. Not surprisingly, the largest improvement was observed for pain reported from the specific joint exposed to surgery (SRM 1.17), but other pain measures also improved (SRM pain VAS 0.43, AIMS2 pain 0.52, SF-36 bodily pain 0.47; fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Magnitude of improvement displayed as standardised response mean across a variety of measures of pain, function and health-related quality of life. *p⩽0.05, **p⩽0.01, ***p⩽0.001 (paired t test for the mean change). AIMS2, arthritis impact measurement scales 2; HAQ, health assessment questionnaire; SF-36, short form 36, BP, bodily pain; GH, general health; MH, mental health; PF, physical function; SF, social function; VT, vitality; VAS, visual analogue scale.

The improvement for physical functioning differed between instruments. The SRM for SF-36 physical was 0.48, but only 0.05 for HAQ and 0.10 for AIMS2 physical. Other dimensions of health also improved, but with small effect sizes, eg, fatigue VAS (0.23) and the SF-36 mental (0.22), vitality (0.24) and social function (0.16) (fig 1).

The pattern of improvement was similar overall after surgical procedures in the lower and upper limb, but changes were, as expected, different for measures that specifically addressed functional aspects in the lower (eg, SF-36 physical, AIMS2 mobility, AIMS2 walking and bending) versus the upper (eg, AIMS2 arms) limbs (table 1).

A larger improvement across all dimensions of health was observed after replacement surgery compared with non-replacement surgery (table 2).

The overall improvement (change) in utility was 0.10 (SRM 0.37) with EQ-5D and 0.03 (SRM 0.27) with SF-6D. Figure 2 illustrates that this improvement was generally captured within 6 months. The other HRQoL instruments showed the same pattern of response as SF-6D and EQ-5D. A large improvement was observed after 3 months and stabilised thereafter.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Mean (95% CI) values of SF-6D (triangle) and EQ-5D (circle) at baseline and 3, 6, 9 and 12 months after surgery.

The gained HRQoL expressed in utilities was larger after replacement surgery than after non-replacement surgery (table 2).

Discussion

Pain is the area of health in which most patients with RA would like to see an improvement.21 The current study demonstrates that a major improvement is observed in patient-reported pain after orthopaedic surgical procedures in patients with inflammatory arthropathies (fig 1). This finding was consistent for procedures in the lower and upper limb and for replacement and non-replacement surgery (tables 1 and 2).

Improvement in the measures of physical functioning was generally smaller than in measures of pain (fig 1). This observation supports the theory that pain in the actual joint is probably the most important indicator of treatment benefit and this observation is of relevance when informing the patients preoperatively about their treatment expectations.

We also observed an improvement in other dimensions of health that are important to patients but not considered as directly related to benefit from surgery. For example, both fatigue and SF-36 vitality, as well as SF-36 mental and social functioning improved more than some of the scales reflecting physical functioning, even if the effect sizes were small or trivial (fig 1). It is known that the level of pain is strongly related to most other self-reported health status measures.22 23 The observed improvements in measures of fatigue, mental and social functioning may thus also reflect reductions in pain intensity. Furthermore, pain has a major impact on the patient global assessment of the disease and we assume that the improvement in global VAS is caused by an improvement in pain.22 This assumption is also supported by a strong correlation between changes in these two dimensions in the current study (r  =  0.64, p<0.01).

A limitation of this study was the lack of other disease activity measures, including joint counts and acute phase reactants. However, inflammatory activity is strongly related both to physical functioning and pain,23 24 and the apparent differences in the magnitude of improvement between these dimensions do not support the belief that the inflammatory activity was considerably changed during follow-up. The rate of non-respondents was similar to the observed rates in mail surveys in the Oslo RA register.25 Patients who underwent replacement surgery had a few percentages more missing values than patients who underwent non-replacement surgery and the oldest patients also had a slightly higher percentage of missing values than younger patients. The missing values were independent of gender and diagnosis. We do not consider the missing values to have any major impact on the results because LOCF was used as a conservative method to handle missing values. A possible influence on the results was probably in the direction of an underestimation of the benefit.

Cost-effectiveness analyses are important in modern medicine as priorities have to be determined between an increasing number of costly therapeutic opportunities. The different magnitude of improvement with EQ-5D and SF-6D was as expected based on the different profiles of these instruments with regard to ceiling and floor effects.17 26

We have previously shown that lower limb function deteriorates more than upper limb function over a 10-year period in patients with RA.27 Therefore, we wanted to compare the outcomes of upper and lower extremity surgery. Many of the replacement surgical procedures in the lower limb are also performed in patients with osteoarthritis.

Replacement surgery was performed in 25.6% of patients, and this frequency is in accordance with previous studies.28 29 Compared with patients who underwent non-replacement surgery, the replacement surgery group reported worse health preoperatively, and a larger improvement was observed during follow-up, although both groups had a significant improvement in most scores (table 2). This preoperative difference in HRQoL between the patients who underwent replacement surgery versus non-replacement surgery may be a bias in the study. Importantly, the mean change in EQ-5D utility was 0.22 in the replacement surgery group and only 0.06 in the non-replacement group. Replacement surgery is more expensive than other procedures, and a larger benefit should therefore also be expected for such procedures, if they can be considered cost-effective. Some studies have shown declining use of orthopaedic surgery in patients with RA inflammatory arthropathies,30 31 32 but also that disability and other HRQoL scores predicted subsequent prosthesis surgery in a cohort of early RA patients.33

Self-administered questionnaires were used for the assessment of all outcomes in the current study. PRO have been shown to provide information similar to many clinical measures and are at least as responsive.6 Furthermore, self-reported questionnaires appear to be a cost-effective and attractive approach for quantitative assessment and monitoring of health status of the individual patient.34 A strength of this study was the use of several instruments capturing the same construct or domain.17 Consistent results across instruments support the robustness of the results. An exception was the results achieved with the instruments measuring physical function, but the scale from SF-36 has a stronger focus on lower limbs than AIMS2 physical and HAQ. However, the large number of instruments represented a burden for the patients as missing values occurred with a frequency of approximately one third at each point of follow-up. Therefore, a conservative analytical approach (LOCF) was used to replace missing values.

Surgical interventions are complicated procedures that can have major and fatal consequences for the patient. The outcome not only depends on the operative technique, the proper selection of the type of surgery and postoperative rehabilitation, but also on the progression of the disease and the state of advancement of pathological changes in other joints. Further research is needed to evaluate the costs and consequences of surgical procedures that are offered to patients with inflammatory arthropathies. RCT are needed, but are difficult to perform.

The current observational study showed that surgical procedures had a major positive impact on pain, but the overall improvement on physical functioning was of similar magnitude as the improvement in psychosocial variables and fatigue. These observations support the belief that pain rather than function should be the major indication for surgical procedures and contributes to the understanding of how patients should be informed about their expectations of surgical procedures.