1 Online Mindfulness vs Progressive muscle relaxation vs a combination of both as a treatment of migraine - empowering clients in their choice of treatment to CONTROL THEIR CHRONIC HEADACHE HSCP Conference Jonathan Egan*, Elena Chepucova, Brian McGuire, & Sinead Conneely NUI GALWAY *Formerly Mater Misericordiae University Hospital

2 Introduction BackgroundPrevalence of active headaches disorders has been estimated at 47 percent for general headaches, ten percent for migraine, and over three percent for chronic daily headache. In the UK alone over 25 million working days per annum are lost, and this is only in relation to migraine type headaches (WHO, 2012). NICE (2012) suggests that 4% of all primary care visits are in relation to headache disorders, this rising to approximately 30% of neurology outpatient appointments.

3 Historical treatmentsRelaxation treatments using both Biofeedback (Blanchard, 1992) training and Progressive Muscle Relaxation (Jacobson, 1938) have tended to be the treatments of choice. These pioneers of the field have been noted in Lambert’s handbook of behaviour and psychotherapy change (2013) as being particularly strict in their definitition of reliable clinical change: needing a sustained 50 percent reduction in headache symptoms. One might argue that this treatment goal might be too high and that smaller reductions in symtoms should be worthwhile.

4 3rd wave Grossman, Nieman, Schmidt, et al. (2004) have reported how mindfulness based stress reduction programmes can result in medium to large treatment sizes for both physical and mental health, however a Cochrane review (Williams, Eccleston & Morley, 2012) of the effectiveness of psychological treatments for chronic pain are in the small to medium effect size range, with psychological acceptance being the main effect and actual disorder disorder treatment being in the small effect size region. More recently, Gard et al. (2012) have reported how regular meditators will have higher brain activation in the insula and the thalamus when presented with a transcutaneous elecrtical stimulus and report the stimulus as less painful when compared between a meditative and resting state.

5 Huguet, A. , McGrath, P. J. , Stinson, J. , Tougas, M. EHuguet, A., McGrath, P. J., Stinson, J., Tougas, M. E., & Doucette, S. (2014). Efficacy of psychological treatment for headaches: an overview of systematic reviews and analysis of potential modifiers of treatment efficacy. The Clinical journal of pain, 30(4), Eighteen reviews met a priori criteria for inclusion. The broad scope of research on efficacy of psychological treatments for Headache is reflected by variation in clinical and methodological characteristics of the reviews. These variations were explored through meta-analysis and subgroup analysis of 41 primary studies and showed that some of these variations, including time of assessment, treatment type, age, HA diagnosis, and study quality, can impact the magnitude of treatment effect.Discussion: There is substantial evidence in favor of psychological treatments for HA management. Further investigation, especially in specific treatments (cognitive-behavioral or autogenic treatment) for HA disorders, is needed.

6 Cochrane Psychological therapies for frequent episodic and chronic tension-type headache in adultsBrian McGuire1,*, Amanda C de C Williams2, John Lynch3, Michael Nicholas4, Stephen Morley5, John Newell6, Ali Asghari7,8Editorial Group: Cochrane Pain, Palliative and Supportive Care GroupPublished Online: 30 SEP 2014DOI:  / CD011309

7 Online Headache & pain interventionsRecent studies - PMR and bio-feedback lead to signficantly reduced headache activity and medication use (Devineni & Blanchard, 2005). Psycho-education, management of nausea, use of a diary, recognising triggers, and learning biofeedback and relaxation resulted in significantly higher self mangement strategies and lower occurrences of migraine (Bomberg et al, 2011). Wells and Loder (2012) described the dichotomy of treatment approaches falling into two main branches: (a) Behavioural; CBT, Stress Management & coping skills, & (b) Mind/Body (meditation, yoga, guided imagery, biofeedback, hypnosis, Tai Chi, Qi Gong, deep breathing exercises and PMR). They suggest tailoring treatments to suit the individual and that a correct treatment suite can result in a 35-50% reduction in migraine and tension type headaches.

8 Russian Headache research society (2012)Virtually no empirical investigations into the efficacy of psychological treatments for chronic headaches within the Russian population over the last decade, with the focus largely being on bio-medical interventions. This introduces the current online study as the first use of a mindfulness based treatment in a Russian speaking sample.

9 Methodology Participants-A power analysis using G Power 3 software indicated a sample size of 69 participants would be required to achieve 80% power, with an effect size of .16 (critical F value 0 3,14, df 2, 66). the actual follow-up sample size was 59 (critical F value= 3.16, df = 2, 56). Recruitment was made via the Russian Headache Society and European Headache Alliance. the study was registered with Current Controlled Trials, registration number ISRCTN

10 definition for inclusionExperiencing chronic headache (The International Classification of Headache Disorders (ICHD-II) definition of chronic headache, defined as “the presence of chronic headache on 15 or more days a month and which had persisted for a period of at least three months”, was utilised in this study [Silberstein et al., 2005]). Russian nationals, over 18 years of age Not involved in other type of meditation or relaxation training Able to provide online consent definition for inclusion

11 Exclusion Criteria The presence of uncontrolled blood pressureSevere mental illness or psychosis which would prevent the ability to understand and participated Secondary headache Having previously received meditation or relaxation training

12 Consort Flow Diagram Mind fulness 31 Comb 28 PMR POST 21 22 Follow-up98 applicants were assessed for eligibility. Eight were initially excluded due to their headache not being defined as chronic (5) or their nationality not being defined as being Russian (3). Leaving 90 eligible for randomisation to Mindfulness body scan intervention or PMR or combined treatment. PRE Mind fulness 31 Comb 28 PMR POST 21 22 Follow-up 19 18

13 Observe subscale of the Kentucky Inventory of Mindfulness Skills (KIMS) (Baer, Smith, & Allen, 2004), the KIMS is a 39-item mindfulness skills assessment inventory . This study utilised only the twelve-item ‘Observing’ subscale, which assesses the extent to which an individual attends to various stimuli. Higher scores indicate more mindfulness. The ‘Observing’ subscale has good internal reliability with a Cronbach’s α of .91 (Baer, et al. 2004), and good content and concurrent validity, correlating with the Mindfulness Attention Awareness Scale (Carlson & Brown, 2005). It also correlates negatively with the Toronto Alexithymia Scale (Baer, et al. 2005). In the present sample, Cronbach’s α was .72. Chronic Pain Acceptance Questionnaire Short Form (CPAQ-8). The CPAQ-8 (Fish, McGuire, Hogan, et al ), comprised of two subscales, Activity Engagement and Pain Willingness, allows for a separate analysis of the two dimensions of chronic pain adjustment. Participants indicate their responses on a scale of 0 (never true) to 6 (always true). Fish, et al. (2010) provide evidence of good internal reliability of the Activity Engagement and Pain Willingness subscales, with Cronbach’s alpha’s of .86, and .85 respectively; and fair-to-good test-retest reliability with interclass correlation coefficients ranging from .50 to .86. The authors validated the scale with a chronic pain sample obtained online. These findings have been widely replicated (Baranoff, Hanrahan, Kapur & Connor, 2013). In the present sample, Cronbach’s α for Activity Engagement was .67, and for Pain Willingness, α was .58. Pain Catastrophizing Scale . The Pain Catastrophizing Scale (Sullivan, Bishop, & Pivik, 1995) is a 13-item measure which evaluates the extent of catastrophic thinking in relation to pain. Responses are indicated on a five – point Likert scale ranging from 0 (“not at all”) to 4 (“all the time”). Higher scores are reflective of more catastrophic thinking. Evidence of discriminant and convergent validity has been provided by Osman, Barrios, Kopper, et al. (1997). The authors also report good internal reliability, Cronbach’s α of .91, and adequate test- retest reliability. These findings have been replicated (Thorn, Day, Burns, et al., 2011). In the present sample, Cronbach’s α was .66. Measures

14 Leeds Dependence Questionnaire (LDQ)The LDQ (Ford, P. 2003), was utilised to evaluate medication dependence. It was modified to include the term ‘analgesic’ instead of the term ‘substance’, in order to make it more suitable for the current sample. The LDQ consists of 10 items designed to measure dependence for a variety of substances, to be sensitive to change over time, and through the range of mild to extreme dependence (Ford, 2003). It has a 0 (“never”) to 3 (“nearly always”) response format. Ford (2003), found the test-retest reliability of the measure to be high (r = .95). Raistrick and Bradshaw (1994), provide evidence indicating that the LDQ has good provide valid and meaningful data (Anderson, Kaldo-Sandstrom, Strom, et al., 2003). In the present sample, Cronbach’s α for anxiety was .52, and for depression, α was .54.

15 Measures Continued Brief Illness Perception Questionnaire (Brief IPQ). The Brief IPQ (Broadbent, Petrie, Main, & Weinman, 2006), is a 9 item scale designed to assess various cognitive and emotional representations of illness. Participants indicate their responses on a 0 to 10 scale. The overall score indicates the degree to which an illness is perceived as threatening or benign, with a higher score indicating a more threatening view of the illness. The Brief IPQ has good test-retest reliability (Broadbent, et al., 2006), with Pearson correlations ranging from .55 to .70. It also has good concurrent validity when compared to the longer Illness Perception Questionnaire – Revised. The authors indicate that the measure can distinguish between different illnesses thereby demonstrating good discriminant validity. In the present sample, Cronbach’s alphas were between .58 and .74. Leeds Dependence Questionnaire (LDQ). The LDQ (Ford, P. 2003), was utilised to evaluate medication dependence. It was modified to include the term ‘analgesic’ instead of the term ‘substance’, in order to make it more suitable for the current sample. The LDQ consists of 10 items designed to measure dependence for a variety of substances, to be sensitive to change over time, and through the range of mild to extreme dependence (Ford, 2003). It has a 0 (“never”) to 3 (“nearly always”) response format. Ford (2003), found the test-retest reliability of the measure to be high (r = .95). Raistrick and Bradshaw (1994), provide evidence indicating that the LDQ has good concurrent and discriminant validity, and high internal reliability with a Cronbach’s alpha of .94. In the present sample, Cronbach’s α was .56.

16 Hospital Anxiety and Depression Scale (HADS)-The HADS (Zigmond & Snaith, 2003), was used to evaluate levels of anxiety and depression. The HADS permits one to do this without allowing physical symptomatology to contaminate the scores (Fish, et al. 2010). It is a 14-item measure, scored 0 to 3. Higher scores suggest higher levels of anxiety and depression. Both subscales have good internal reliability; Cronbach’s alpha for the Anxiety subscale ranges from .68 to .93, while for the Depression subscale the range is between .67 and .90 (Bjelland, Dahl, Haug, et al., 2002). Internet administration of the HADS among a sample of tinnitus patients has been shown to provide valid and meaningful data (Anderson, Kaldo-Sandstrom, Strom, et al., 2003). In the present sample, Cronbach’s α for anxiety was .52, and for depression, α was .54.

17 Online Participant recruitment, data collection and training, were conducted online. Potential participants who visited the website and wished to participate, received the combined online participant information sheet and informed consent form. Consenting eligible participants proceeded to respond to the full survey (Survey Monkey TM). Upon completion, participants received login details which allowed for access to one of three groups: MM, PMR, or combined MM and PMR. Participants were blind to treatment allocation until they accessed the programme for the first time. Upon consenting, participants were assigned a unique identification number, and were randomly allocated to one of three experimental groups. The researcher was not blind to treatment allocation. A random integer sequence was generated on the computer software programme Graphpad©, allowing for a list of participants’ individual identification numbers to be randomly associated with a number from 1 to 3, specifying group allocation. Exercises for MM were provided by presenter Dr. Jonathan Egan, while PMR exercises were adopted from Carr and McNulty’s (2006) scripts for PMR in the Handbook of Clinical Psychology. Audio recordings were conducted by a native Russian speaker, and embedded in the website. Participants’ frequency and duration of usage was tracked by the software programme StatCounter©.

18 Upon consenting, participants were assigned a unique identification number, and were randomly allocated to one of three experimental groups. The researcher was not blind to treatment allocation. A random integer sequence was generated on the computer software programme Graphpad©, allowing for a list of participants’ individual identification numbers to be randomly associated with a number from 1 to 3, specifying group allocation. Exercises for MM were provided by presenter Dr. Jonathan Egan, while PMR exercises were adopted from Carr and McNulty’s (2006) scripts for PMR in the Handbook of Clinical Psychology. Audio recordings were conducted by a native Russian speaker, and embedded in the website. Participants’ frequency and duration of usage was tracked by the software programme StatCounter©.

19 RESULTS Data was analysed by intention-to-treat. Multiple imputation was utilised to deal with missing data due to drop out. All variables were tested for normality. Measures were analysed using a series of 3 x 3 mixed factorial analyses of variance (ANOVA), with between subject factor Group, and a repeated-measures within subject factor Time. Maulchy’s test of sphericity was violated on all measures, therefore the Hunyh-Feldt correction is reported. Categorical demographical information was analysed using frequencies. Effect sizes obtained in this study are presented in the following table.

20 Group Descriptive DATAPreliminary Analysis- A randomisation check by multivariate analysis of variance revealed that all three groups were comparable in age, gender, stage of menstrual cycle, type of headache, how long they had suffered from chronic headache, and scores on all of the dependent variables (Wilks ʌ = .471, F (48, 124) = 1.182, p = .230).

21 Results Continued Bonferroni post-hoc tests revealed that participants in the MM (M = 35.54, SD = 6.16), and the Combined (M = , SD = 6.16) groups, reported significantly higher mindfulness skills than the PMR group (M = 27.97, SD = 7.44).

22 Pain severity

23 Pain interference

24 ACtivity Engagement

25 Pain Willingness

26 Medication Dependence

27 Medication Dependence

28 Catastrophising

29 Pain Catastrophising

30 Anxiety

31 Depression

32 Illness Perception

33 Summary of effect sizes

34 Discussion Limitations include small sample size and level of attrition Ease of techniques suggests even a small to medium effect size which has an effect for six weeks, it should not be ruled out as a valid treatment for the short to medium term. HSCPS need to consider online studies to identify factors related to prolonged and lasting benefits, as well as to inform the patient of best tailored treatment for them.