Patients’ perspective of the effectiveness and acceptability of pharmacological and non-pharmacological treatments of fibromyalgia

2.1 Questionnaire development

The questionnaire was developed from the outcomes of a pilot study of individual qualitative interviews [28] and a review of existing literature. A self-administered anonymous questionnaire was conducted using Qualtrics [33] that asked participants to provide quantitative and qualitative data about a variety of areas including demographics, condition profile, condition impact, treatment usage and treatment outcomes. Thus, the structure of the survey included items related to the physical and psychological symptoms and management. Members of the fibromyalgia research team at Sheffield Hallam University reviewed the structure, questions and ease of completion to ensure the survey was comprehensive, balanced and user-friendly. This paper focuses on the quantitative data on treatment outcomes from the survey.

This research project received ethical approval from the Sheffield Hallam University Research Ethics Committee.

2.2 Recruitment and data collection

Fibromyalgia charities and networks, and affiliated member support groups within the UK posted the survey online with additional notification through emails and on Facebook pages from January 2016 to April 2016. To take part in the survey, participants were required to be over 18 years of age and to have received an official diagnosis of fibromyalgia. Participants were informed about the voluntary and anonymous nature of the survey, the information that would be collected and the expected time for survey completion. Completion and submission of the survey was taken as implied consent and participants were free to withdraw their data up to 2 weeks after participation. To avoid potential fatigue, participants had the option of taking breaks during survey completion. Average completion time was 1 h 15 min.

2.3 Measures

The first section asked participants to provide demographic information including age, gender, employment status, relationship status and qualifications. The next section focused on treatment-specific information with participants asked to state whether they had tried, had not tried, did not know about or would not try identified treatment options. The survey included a range of commonly used pharmacological and non-pharmacological treatments. Participants were asked to rate their personal experience of tried treatments on the basis of the effectiveness and severity of side effects on 11-point Likert scales (effectiveness of treatment 0=not effective to 10=very effective and side effects 0=none to 10=terrible). Further questions required participants to select one symptom [from a choice of nerve pain, fatigue, brain fog, muscular pain, bone/joint pain, anxiety, depression, irritable bowel syndrome (IBS)/gastric issues] that the treatment was most effective in reducing and select which side effect (from a choice of feeling nauseous, feeling tired, feeling disorientated, made pain worse, worse mood, weight gain, dry mouth, anxiety, incontinence, constipation, pain during treatment or none) gave the worst experience.

Participants completed the “Sensory Reactivity” and “Pain and Consequences” subscales from the Comprehensive Rating Scale for Fibromyalgia Symptomatology (CRSFS) [34]. Participants rated whether they had experienced a range of symptoms never, once, on several days, more than half of days or almost every day over the previous 2 weeks causing discomfort or affecting daily life. Sensory reactivity items focused on joint discomfort, pain, stiffness, sensation changes, night cramps and tingling sensations. Pain and consequences items assessed sleep quality, neck and back pain, headache and physical and general tiredness and were scored as follows: never/once (0), several days (1), more than half of the days (2) and almost every day (3). Scores of the subscales were totalled giving a possible range of 0–54 for symptoms of sensory reactivity and 0–63 for symptoms of pain and consequences. Higher scores indicate greater sensory reactivity and pain and consequences. The CRSFS Sensory Reactivity (Cronbach’s alpha=0.89) and Pain and Consequences (Cronbach’s alpha=0.87) subscales had good reliability.

The Hospital Anxiety and Depression Scale (HADS) [35] assessed state anxiety and depression with seven items scored between zero and three on each subscale. “I feel tense or wound up” is an example from the anxiety subscale and “I feel cheerful” is an example from the depression subscale. Scores of 0–7 were categorised as normal, 8–10 as mild, 11–15 as moderate and 16–21 as severe [35]. The Anxiety (Cronbach’s alpha=0.84) and Depression (Cronbach’s alpha=0.83) subscales had good reliability.

2.4 Analyses

Summary statistics were calculated as means and standard deviations (SD) for continuous variables and frequency (%) for categorical variables. Data were ranked by decreasing effectiveness scores. The range of effectiveness and side effects ratings for pharmacological and non-pharmacological treatments were compared by Mann Whitney U-test with a significance level of 0.05. A treatment acceptance ratio was calculated for each intervention from the mean effectiveness value and the mean side effects value. The range of acceptance ratios for the pharmacological and non-pharmacological treatments were compared by Mann Whitney U-test with a significance level of 0.05. Mann Whitney U-tests were used because these analyses were based on ranked data.

The participant-reported outcomes were evaluated within the different classes of treatments to identify the preferred therapies. Between group ANOVAs comparing the effectiveness and severity of side effects were reported for treatment groups with three or more treatments. Hochberg’s GT2 post hoc tests were reported because these tests are recommended when sample sizes are different [36] and there is some variation in number of participants who tried different treatments. t-Tests are reported when there were two treatment types in a group. The significance level was set at 0.05. Parametric analyses were conducted because box plots and skewness statistics indicated that the data were normally distributed.

3.1 Participant demographics

Nine hundred and forty-one participants started the survey with 76% of participants completing 50% or more of the survey. t-Tests were conducted to compare participants who had completed the entire survey with participants who had not completed the entire survey in terms of age, CRSFS sensory reactivity, CRSFS Pain and Consequences and HADS Anxiety and Depression scores. Participants who completed the whole survey were significantly older (M=48.83, SD=12.24) than participants who did not complete the whole survey (M=46.57, SD=11.84, t (936)=2.88, p=0.004). Participants who finished the survey reported significantly lower CRSFS sensory reactivity (M=34.99, SD=10.63) relative to participants who did not finish the survey (M=36.78, SD=10.27, t (906)=2.58, p=0.010). Participants who completed the survey reported significantly lower depression scores (M=10.29, SD=4.35) compared to those who did not complete the whole survey (M=10.97, SD=4.26, t (931)=2.42, p=0.016). There were no differences in terms of CRSFS Pain and Consequences (t (890)=1.62, p=0.106) or Anxiety (t (931)=2.42, p=0.051).

Participant demographic data are presented in Table 1. The demographic characteristics indicated that the respondents were 94.8% white females with a mean age of 47 years. Fifty-two per cent of participants were married or in a partnership, 40.1% were in employment and 30.5% had a university education. The time since diagnosis ranged from 0 to 37 years with a mean of 5 years.

3.2 Use and knowledge of treatments

Participant engagement with UK-licensed pharmacological and non-pharmacological treatments are presented in Tables 2 and 3. Six of the considered treatments had been tried by 75% or more of the participants: paracetamol (92.5%), heat pad/hot water bottle (81.8%), codeine (79.3%), ibuprofen (77.6%), walking/running (76.1%) and amitriptyline (75.9%). Only five of the treatments were identified by greater than 20% of participants as “would not try”: cycling (55%), gym exercise (39.5%), steroid tablets (28.6%), morphine (28.3%) and fluoxetine (24.4%). The participants were aware of the majority of the non-pharmacological treatments considered, other than colpermine which 45.3% of participants did not know (Table 3). In contrast, Table 2 shows that a third of the pharmacological treatments considered were not known by greater than 50% of participants: Audmonal (alverine) (73%), nefopam (72.2%), nortriptyline (58.1%), fentanyl/Butrans (buprenorphine) patches (57.2%), sertraline (55.3%), ranitidine (54.6%) and duloxetine (51.7%).

3.3 Effectiveness and side effects of treatments

Tables 4 and 5 show the mean effectiveness and side effects ratings of pharmacological and non-pharmacological treatments. The range of mean effectiveness scores, rated on an 11-point Likert scale for pharmacological (2.54–6.85) and non-pharmacological treatments (3.54–6.33) were similar (U=197.5, p>0.05) (Tables 4 and 5). The predominant symptoms identified as gaining greatest relief, pain, depression and anxiety, again were similar within both types of treatment. In contrast, a significantly lower range of mean side effects scores was reported for non-pharmacological treatments (0.62–6.13) relative to pharmacological treatments (1.17–6.34) (U=132, p=0.027). The predominant side effects identified with pharmacological treatments were nausea and disorientation, whilst with non-pharmacological treatments were pain and tiredness (Tables 4 and 5).

3.4 Treatment classes

3.5 Participant acceptability of treatments

The relationship between the effectiveness of a treatment and the side effects was determined as an assessment of participant-rated acceptance. An acceptability ratio greater than 1 indicated that the benefits outweighed limitations as assessed from the participants’ perspective. The acceptability ratio for pharmacological treatments range was 0.67–3.92 with nine of 21 interventions giving a value greater than 1 (Table 6). For non-pharmacological treatments, the acceptability ratio range was 0.60–9.65 with 15 of 21 interventions giving a value of greater than 1 (Table 7). The acceptability ratio range for non-pharmacological treatments was significantly higher than that for pharmacological treatments (U=141.5, p=0.048).

4.1 Previous research

Survey results supported a previous fibromyalgia survey that found non-pharmacological treatments were more effective at symptom relief than pharmacological treatments with worst side effects reported for strong opioids and gabapentinoids [32]. The current survey results showed limited effectiveness of pharmacological treatments including analgesics, gabapentinoids and antidepressants, supporting the EULAR guidelines that suggested weak evidence for the efficacy of pharmacological treatments [21]. A review of pharmacological therapies in fibromyalgia highlighted the importance of tailoring drug therapy based on key symptoms other than pain, psychological and physical comorbidities and the importance of problematic side effects to the individual e.g. weight gain [15]. Research suggests there are subgroups of fibromyalgia symptoms based on personality with cluster 1 characterised by higher neuroticism and lower extraversion, remaining more anxious and depressed during treatment, compared to cluster 2 [37]. The clusters should be included in the design of future treatments and clinicians should be aware of cluster types during patient interaction. An alternative subgroup categorisation in fibromyalgia patients is based on (1) functional, (2) dysfunctional and (3) highly dysfunctional and distressed profiles [13].

4.2 Activity

The EULAR and Canadian guidelines strongly recommended exercise because of pain reduction, wellbeing and low cost [21], [24]. Self-report data from the survey, however, showed acceptability for activity types (gym exercise, cycling, walking/running, swimming and exercise plans) was less than 1 indicating that undesirable effects outweighed effectiveness. It is possible that benefits of activity are only evident when participants take part in regular activity. More research is needed to determine the recommended intensity and frequency of activity for benefits of symptom relief to be achieved [38]. High attrition rates and participants not adhering to activity, however, are problems associated with activity interventions [39]. Comparing the different types of activity included in the survey, swimming was more effective than exercise plans and participants reported lower side effects compared to gym exercise, cycling, walking/running and exercise plans. Swimming may be beneficial to pain because hydrostatic pressure and buoyancy may reduce muscle spasm and increase the pain threshold of nerve endings resulting in decreased pain sensation [40].

4.3 Physiotherapy

Enhanced pain was a commonly reported side effect for activity and physiotherapy. A systematic review identified increased pain during activity as a barrier to physiotherapy adherence in musculoskeletal outpatients, together with low levels of physical activity at the start of therapy, low adherence to exercise during therapy, low self-efficacy, depression, anxiety, helplessness, poor social support and greater perceived barriers to exercise [41]. Simple analgesics, heat or ice with passive physiotherapy (e.g. acupuncture or manual therapy) may reduce pain allowing the use of more active treatments [42]. Physiotherapists should aim to change maladaptive pain beliefs when patients believe that pain experienced during physiotherapy is harmful [41]. Patients with fibromyalgia would benefit from more specialist physiotherapists and psychologists in the healthcare system and integrated interdisciplinary care [43].

4.4 Dietary changes

Participants reported that diet change helped to alleviate gastric symptoms and had low side effects. Research shows that for subgroups of patients, dietary changes, such as gluten-free diet, were associated with improvements in fibromyalgia symptoms [44]. Noxious macromolecules may trigger release of mast cell mediators and immune system activation, resulting in pain [45] so that reduction in noxious macromolecules through diet change may improve pain symptoms. Dietary changes are not included in current guidelines, although it has been suggested that nutrition has the potential to be recommended as an effective therapy following more research [46]. Fibromyalgia treatment should include specific dietary interventions, weight loss strategies, nutritional education and tailored nutritional supplements [47].

4.5 Psychological therapy

The results of this survey showed that mindfulness was more effective than CBT and pain management courses with 33% of participants who had tried mindfulness reporting it alleviated anxiety. This contrasts with the Canadian guidelines that highly recommend CBT [24]. One meta-analysis reported small short-term effects of mindfulness-based stress reduction on pain and quality of life in people with fibromyalgia, although there was only a weak recommendation due to the small number of studies and low-quality evidence [48]. Developing coping skills is important in treating fibromyalgia [24] and mindfulness encourages an accepting pain coping style, rather than reducing pain [48]. Twenty-six per cent of participants in the current sample reported they did not know about mindfulness as a treatment for fibromyalgia indicating that mindfulness should be more widely recommended with online resources and courses advertised to people with fibromyalgia.

4.6 Heat pad treatment

Heat pad treatment had the highest acceptability ratio with high effectiveness and low side effects reported. Heat treatment (i.e. thermotherapy) is a very old method of pain relief [49]. There is evidence that participants who received whole-body hyperthermia showed a significant reduction in pain compared to a control group [50]. In the current survey, 74% of participants who had tried a heat pad reported that it alleviated muscular pain and 94% reported no side effects. This supports the findings of a previous online survey [31] that reported heat pad was the fourth most effective treatment in an internet survey of participants with fibromyalgia (following prescription sleep medication, prescription pain medication and resting). Eleven per cent of participants reported they had not tried this treatment and only 3.8% of participants reported they would not try using a heat pad. Healthcare professionals should discuss using heat pads with people who have fibromyalgia because this treatment was found to be highly acceptable, easy to administer and inexpensive. While the exact mechanism of heat and pain reduction is unclear [51], sedative effects on sensory nerve endings [52] and capillary dilation [53] are possible explanations.

4.7 Distraction therapy

Distraction therapy through hobbies was another treatment that participants rated as effective with a low number of side effects. An explanation for this finding is that attention is diverted away from symptoms. A systematic review of distraction-based treatments noted that it is important to examine predictors of treatment outcomes, including personality, so that pain management strategies can be individualised to the patient [54].

4.8 Strengths and limitations

Strengths of this study are the relatively large community participant sample exhibiting profiles typically expected of patients with fibromyalgia and the patients’ perspective reflecting a real-world setting. In addition, the survey included effectiveness and side effects of treatments, important factors in treatment adherence and acceptance. Furthermore, the survey included a range of treatments: pharmacological, activity, psychological, physical and alternative therapies. Strengths and limitations inherent to online survey research must also be noted and considered in the interpretation of the results. Strengths include access to specific populations, the collection of a large amount of data quickly and low cost, although self-selection bias is a potential limitation because some individuals may be more likely to respond than others [55].

Limitations of the sample should be acknowledged. For example, the sample contained a lower percentage of participants who had a university education (30.5%) compared to 40% in the general population [56], possibly indicating sampling bias. Furthermore, participants who completed the entire survey were significantly older and reported significantly lower sensory reactivity and depression scores compared to participants who did not complete all the survey. Therefore the current sample may not accurately reflect the wider demographic of fibromyalgia patients. The survey was advertised by fibromyalgia organisations and on social media, thereby participants had Internet access and an active interest in the topic. The use of patient advocacy groups and websites to access respondents could lead to over-representation by individuals seeking support and fellowship because of experienced problems. Although participants were required to have received an official diagnosis of fibromyalgia, the diagnostic method was not determined and could not be confirmed, and fibromyalgia severity was not established. A further limitation was that participants were not asked about current treatments and the past tense emphasis of the questions (e.g. have tried) may have led to recall bias of over- or underestimation of patient reports. It is possible that participants were using more than one treatment, following guidelines that suggest a multimodal treatment approach. The survey did not include dose, intensity or method of administration of treatments and these may have affected effectiveness and side effects ratings. Therefore, survey results show patients’ perspectives of effects and side effects of the doses and duration of drug treatments commonly used for fibromyalgia.

In conclusion, the main survey findings were that participants reported lower mean side effects and higher mean acceptability, although mean effectiveness was similar, for non-pharmacological treatments compared to pharmacological treatments. Participants were not aware of some treatment possibilities indicating that patient education is important for highlighting the range of available treatments allowing optimal informed treatment decision making with their healthcare professional.