The Maintaining Musculoskeletal Health Study

Chronic widespread pain (CWP), the cardinal feature of fibromyalgia, is associated with lost work productivity, psychological ill health, and poor quality of life. It is one of the most common reasons for referral to a rheumatologist. The cost of CWP is high in terms of both individual, societal and health costs: for example, in the United States, mean per-patient costs (including pain and non-pain-related medication, physician consultations, tests and procedures, and emergency department visits) in the 6 months following a new diagnosis of fibromyalgia have been reported as $3481, comparable to patients with rheumatoid arthritis but resulting in worse quality of life. Current guidelines recommend pharmacological, physical, and psychological therapies although the importance attributed to individual therapies is inconsistent. There is good evidence for musculoskeletal pain conditions generally that the longer the duration of symptoms, the less likely that symptoms are to improve, including with specific interventions. This is particularly so for CWP which, once developed, is challenging to manage and effect improvement.

A systematic review and meta-analysis of randomised controlled trials (RCT) of Cognitive Behaviour Therapy (CBT) for patients with fibromyalgia concluded that CBT improves coping with pain, reduces depressed mood and healthcare-seeking behaviour in such patients. The delivery of CBT by telephone has been shown to be effective, acceptable and accessible. The MUSICIAN study, which the investigators have recently concluded, tested telephone delivered CBT (tCBT) and/or exercise for patients with chronic widespread pain consulting to their GP. Three months after the end of therapy, both interventions resulted in significantly better primary outcome measures (patient global health) than treatment as usual, but there was no significant additional benefit of receiving both interventions. Recent analyses have demonstrated these benefits are maintained 2 years after the end of therapy. The investigators have conducted a comprehensive literature review with the aim of identifying randomised trials which had the aim of preventing the onset either of CWP or fibromyalgia. This review did not identify any such published trials. Further, a search of 11 international clinical trials registers/databases (including US, UK, Europe, Australia/New Zealand, Japan), undertaken in Autumn 2013 did not identify any ongoing trial with the aim of preventing the onset of CWP (or fibromyalgia).

There are several reasons why it may be desirable to try to prevent CWP onset, namely that the majority of CWP patients do not have important symptom improvement with current management (even within trials). Prediction models from epidemiological studies have been developed to identify "high risk" patients, which makes such an approach feasible. Research using the General Practice Research Database has demonstrated that prior to receiving a diagnosis of fibromyalgia in primary care, persons have a long-term prior history of consultation with symptoms. Although this will be the first prevention trial in this area, the concept of prevention using CBT has been addressed in musculoskeletal disorders with respect to intervention in neck pain and low back pain before people become patients and in mental disorders.

The investigators have conducted prospective epidemiological studies which have demonstrated that it is possible to identify "high risk" groups. In the first study, a high risk group for CWP onset was identified on the basis of two factors: somatic awareness (using the Somatic Symptom Scale) and illness behaviour (using the Illness Behaviour Score). This was replicated in a second study conducted by the applicants. These "aetiological models" excluded pain and therefore the investigators have re-analysed data from the latter study (also considering pain status) to identify the best predictors of onset and which results in a model more suitable for use in prevention studies. The resulting "at risk model" requires regional pain and two of the following: maladaptive behavioural response to illness), a high number of somatic symptoms, and sleep disturbance. In the second "validation" study, from a population of 2,374 persons without CWP, 653 satisfied the definition of "high risk of CWP" of whom 139 had developed CWP twelve months later (that is a Positive Predictive value of 21.3%). Amongst persons not deemed to be at high risk (n=1721), 77 developed CWP which is a Negative Predictive Value of 95.5%.

The investigators have previously shown short and long-term effectiveness of tCBT for CWP (compared to usual care). Specifically this demonstrated sustained improvement in patient global assessment of change, reduced psychological distress, fear of movement and reliance on passive coping styles. Secondly the investigators have developed and refined statistical models which identify persons at high risk for the future development of CWP. The investigators therefore now propose a study to test whether tCBT can reduce the risk of CWP onset amongst those at high risk.

Three health boards in Scotland will be research sites for the study. The three health boards are NHS Grampian, NHS Highland, and NHS Greater Glasgow and Clyde. The study will require the involvement of 7 or 8 equivalent general practices. The investigators will mail a randomly selected sample of adults aged 25 years and over registered with participating general practices in the study areas. The Scottish Primary Care Research Network (SPCRN) will be involved in recruitment of patients to this study from primary care. SPCRN staff provide GP practices with an expert service to undertake searches of their electronic databases to identify a random sample of potentially eligible patients and prepare the ethically approved letters to be sent out. Searches are undertaken at each GP practice before the screening survey questionnaires are sent out by Health Informatics Centre Services in Dundee on behalf of the practice. Patients will return completed survey questionnaires to the research team at the University of Aberdeen where responses will be assessed by the research team for eligibility and patients sent invitation letters if eligible.

The "screening questionnaire" will determine whether a) respondents meet the study eligibility criteria and b) respondents would be willing to be contacted again regarding a treatment trial for "musculoskeletal health". The questionnaire will include:

• Pain assessed by specific questions on the experience of pain, consultation and body manikins (which will provide site and also allow us to exclude those who already have chronic widespread pain).
• Illness Behaviour Scale
• Somatic Symptoms Scale (excluding pain items)
• Sleep Problem Scale
• Quality of Life and Wellbeing
• General Health Questionnaire
• Chalder Fatigue Scale A list of eligible patients will be provided to the general practitioner in advance, with the option of indicating any as unsuitable for the study. Patients would then be sent information about the study and subsequently contacted by a member of the research team by telephone and, if appropriate, consented and recruited into the trial.

Included in the mailing to eligible patients will be an information sheet, consent form, and a best-time-to-call slip. Once a patient has returned a signed consent form and best-time-to-call slip, a member of the research team will phone them. The researcher will read out from a script giving information about the study, and the patient will have the opportunity then to ask questions about the study. If the patient consents to participate the participant will be recruited to the study and randomised to one of the arms of the trial.

Follow-up questionnaires will be mailed to participants at 3, 12 and 24 months after the treatment start date (for participants in the active treatment group) or dummy treatment start date (for those in usual care). Instruments included in the follow-up questionnaires will be the same as in the screening survey questionnaire. Additionally, follow-up questionnaires will include the Patient Global Impression of Change, and questions on health care usage.

Participants will have the option to withdraw from the treatment or the study at any time. Those withdrawing from the treatment will continue to be sent follow-up questionnaires unless the participant requests not to receive them. Failure of any participant to complete a follow-up questionnaire at any particular timepoint will not be counted as a withdrawal unless the participant requests not to receive any further follow-ups.

Our previous longitudinal study of onset of CWP (and subsequent replication) has suggested that 21% of "high risk" persons identified will develop CWP over the course of the next twelve months. Our previous data is based on persons with pain and at least 2 out of 3 other "risk factors". There are no published studies of prevention of CWP on which to base our measure of effect. However in the MUSICIAN study some subjects, although reporting CWP at the screening survey, no longer had CWP at the enrolment interview. These participants were however still eligible to take part, provided the participant had regional pain. Therefore those subjects with regional pain provide a sub-population on which to base the likely effects of the tCBT. Amongst such subjects, those who received tCBT had a reduced odds of having CWP at the end of the study OR 0.5 95% CI (0.2-1.4) compared to those in usual care.

Thus the study is powered on the ability of the current study to reduce the onset of CWP from 21% to 12%, with 90% power and a 5% significance level. The investigators further assume, based on prior data, that 75% of persons allocated to the tCBT arm will be adherent to the intervention, and that 80% of all subjects will return the follow-up questionnaires to assess outcome.

Accordingly the investigators require 473 subjects per arm that is a total of 946 subjects recruited. In MUSICIAN exactly 50% of those found eligible and willing to consider taking part ultimately were randomised. A previous trial of a cognitive-behavioural intervention to prevent chronic pain found that 36% of patients identified as eligible ended up being recruited to the study. If 80% of eligible patients agreed to be contacted about taking part, this equates to 45% of those eligible and willing to consider taking part being randomised - higher numbers for a clinical trial of CWP reflect the fact that this is a prevention trial rather than a treatment trial and may be less attractive to potential participants. Thus the investigators aim to find a total of 2102 subjects who are eligible and willing to consider taking part. Assuming a participation rate to the survey of 30%, that 1 in 4 people will be "at risk", and (using data from MUSICIAN) that 80% of people who return a questionnaire agree to consider taking part, the investigators require to survey 35 037 persons.

A pre-defined statistical analysis plan will be developed and signed off by the trial steering committee before undertaking any data analysis. Comparison between arms will be on an intention-to-treat basis (main analysis) with a per protocol sensitivity analysis.

Characteristics of the study participants in the two treatment arms will be described using simple summary statistics. Descriptive statistics will include mean and standard deviation for normally distributed continuous data, median and inter-quartile range for skewed continuous data and count and percentage for categorical data. No formal statistical comparisons will be made between baseline characteristics. Primary and secondary outcomes will be described at the three follow-up times: 3, 12 and 24 months, using appropriate summary statistics.