Participants
Men and women aged 18 years or older with RA (meeting American College of Rheumatology criteria), in functional classes I, II or III [7] and attending rheumatology clinics at Selly Oak Hospital in Birmingham were invited to participate either on referral for physiotherapy or by invitation in clinics or by mail. For inclusion, participants gave informed consent in writing and were required to understand and follow simple instructions in English. Patients needed to be on stable doses of disease modifying anti-rheumatic drugs (DMARDs) for 6 weeks and NSAIDs for 2 weeks before entry. Injections with corticosteroids in the 4 weeks before study entry were not permitted but drug changes and injections were permitted during the study to reflect the pragmatic nature of our study. Data on drug therapy was collected at baseline and when final assessments were made.
Patients who had had surgery in the 3 months before study entry or those who had surgery planned were excluded. Other exclusions were patients who had received physiotherapy or hydrotherapy in the 6 months preceding entry: this was done to avoid possible carry-over effects of previous therapy. Patients with known chlorine sensitivity, an infected open wound, poorly controlled epilepsy, hypertension, diabetes, incontinence of faeces, and a fear of water precluding hydrotherapy were excluded. Also excluded were: pregnant women; patients with co-morbid conditions which, in the opinion of the assessing physiotherapist, prevented safe use of hydrotherapy; known carriers of methicillin resistant staphylococcus aureus in the upper respiratory tract; and those who weighed more than 102 kg. The latter because of the safety procedures established for our pool.
A single senior physiotherapist assessed study eligibility in all cases. The Research and Development and Ethics committees of University Hospital Birmingham NHS Foundation Trust gave ethical approval for this study.
Randomisation
Patients were randomised to hydrotherapy or land exercises using sealed opaque envelopes indicating treatment allocation. Randomisation envelopes were prepared at study inception and random number sequence was obtained by flipping a virtual coin [8]. A research assistant not involved in the conduct of the study randomised patients, allocated treatment and collected key data.
Interventions
Participants, in groups, received either a weekly 30-minute session of hydrotherapy (at 35°C) or land based exercises for 6 weeks. Participants were asked to attend weekly but, allowing for sickness and leave, were allowed to default for up to three sessions as long as 6 sessions in all were completed. Those defaulting more than 3 sessions were considered to be treatment failures but were followed up, if possible, to obtain study data. Written instructions on home exercises were provided at the outset to all patients. Patients were not required to do exercises between treatment sessions but could do so if they chose. The physiotherapist who had assessed eligibility supervised treatment in both groups, with support from two other senior physiotherapists. A rolling treatment programme was operated for convenience, therefore group sizes varied between 1 to 4 for hydrotherapy and 1 to 6 for land exercises. The size of our hydrotherapy pool only permitted treatment of 4 patients at any one time. However group sizes in both treatment arms were comparable.
The exercise content in each group was similar and exercises were tailored to each individual's ability. Participants warmed up, by mobilising and stretching. The core exercises, repeated each week, focussed on joint mobility, muscle strength and functional activities. The degree of difficulty was reviewed weekly to ensure each participant made progress at their individual pace. A cool down phase concluded each session. Functional limitations of participants were considered at all times.
Outcome measures
A research assistant blind to treatment allocation assessed key outcome measures. The primary outcome was self-rated overall effect of treatment, measured once; on the day treatment was completed. This validated measure has previously been used in clinical trials of exercise in fibromyalgia and chronic fatigue syndrome [9–11]. Effect of treatment is measured as change on a 7-point scale ranging from 1 (very much worse) to 7 (very much better). Patients were asked: 'please indicate how you feel after your treatment'. The outcome was dichotomised so that participants scoring 6 (much better) or 7 were regarded as responders and others as non-responders [10].
Secondary outcomes were collected at baseline, on the day of the last treatment session and 3 months post treatment. These included: pain, assessed on a 10 cm visual analogue scale (VAS), where 0 cm represented no pain; physical function assessed with the health assessment questionnaire (HAQ); ten metre walk speed, an outcome widely used to assess lower limb function in neurology [12] and piloted previously in our unit [13]; and a EuroQol-5D (EQ-5D) valuation questionnaire comprising a self report of health related quality of life (EQ-VAS) and a health status valuation (EQ-5D index or utility score) [14, 15].
Statistical considerations and analyses
Sixty patients were needed in each treatment group to detect a 28% difference between the two groups in the primary outcome measure with a two-sided significance level of 0.05, and a power of 80%, allowing for a 10% drop-out rate. This estimate was based on a pilot study showing that 73% of patients receiving hydrotherapy experienced an increased sense of well-being [13]. We assumed that around 45% of patients with RA undertaking land exercises would feel 'much better' or 'very much better' based on trials of exercise in other conditions [11]. Binary outcomes were analysed by Fisher's exact test and continuous variables by Wilcoxon signed rank tests for within group comparisons or by Mann-Whitney tests for between group comparisons. Data analyses were done according to the principles of intention to treat.