Chronic Headache Education and Self-management Study (CHESS) – a mixed method feasibility study to inform the design of a randomised controlled trial

Patient and public involvement (PPI) was built into the key stages of the feasibility study to ensure that the research focused on issues that were important and relevant to patients and the public [14]. At the start of the study we established a lay advisory group of people with chronic headache to work with collaboratively. We identified members of the group from Universities/User Teaching and Research Action Partnership (UNTRAP) at the University of Warwick and sent out an advert to our three partner headache groups: Migraine Trust, Migraine Action and National Migraine Centre (In 2018 Migraine Action merged with Migraine Trust). The CHESS Lay Advisory Group specifically supported our application for ethical approval for the study, development of the headache classification interview, development of the study intervention and the choice of patient reported outcome measures.

The aim of this part of the study was to test the feasibility of our recruitment procedures and recruit a sample of participants to test the telephone headache classification interview, to pilot the education and self-management intervention, and to test the feasibility and the outcomes measures.

We developed a telephone headache classification interview for use by a non-headache specialist to classify chronic headache types for reporting and analysis purposes and that could also be used as part of the study intervention to allow targeted treatment and advice. In brief, we did a systematic literature review to identify any existing tools used to classify or diagnose different headache types which was presented to delegates at a headache classification consensus conference attended by headache specialists and people with chronic headache [6]. At the consensus conference delegates agreed what were the important questions to include in the classification interview. The classification interview was not intended to have a rigid interview structure or set questions, instead the person conducting the interview was encouraged to use a logic model to inform their clinical reasoning and decision-making.

We aimed to test the feasibility of training nurses to use the classification interview to classify chronic headache disorders and test the reliability of the tool. To validate the classification interview we trained six nurses, all non-headache experts, to conduct the interviews. The training included a one-day workshop delivered by a neurologist specialised in headache plus time with a member of the study team to practice classification interviews using mock scenarios and a training manual.

Participants from the feasibility study were interviewed first by the nurse and later by a doctor from the National Migraine Centre. The doctor classification was the assumed ‘gold standard’. Participants were classified into: definite chronic migraine, probable chronic migraine or chronic TTH (with or without medication overuse) or ‘other’ headache type (other chronic primary headache or suspected secondary headache). We measured level of agreement between the classifications by nurses and doctors by using simple kappa statistics and prevalence-adjusted bias-adjusted kappa (PABAK).

The development and evaluation of the telephone headache classification interview is described in detail elsewhere [13].

We developed the education and self-management intervention using the Medical Research Council (MRC) framework for complex interventions [7]. Development was informed by three systematic reviews 1.prognostic factors in chronic headache [18], 2. education and self-management interventions for chronic headache [19] and 3. the lived experiences of chronic headache. We drew from the experience of a previously tested self-management intervention for chronic pain [20] and we did qualitative interviews with people with chronic headache to inform the intervention design. The qualitative interviews were with members of the charity Migraine Action to gain their views on what was important to include in the education and self-management intervention. We held a collaborative intervention design meeting, attended by headache specialist clinicians, headache charity representatives, lay people with chronic headache, psychologists, and researchers.

The education and self-management intervention was intended to be delivered in a group format (8–10 per group) facilitated by a nurse and a lay person (with chronic headache). Topics included in the intervention were: understanding headache mechanisms, medication management, mood and headache, recognising unhelpful thought patterns and behaviours, stress management, sleep management, communication and mindfulness. The two and a half day programme used a range of methods including: group discussions, sharing narratives and experiences, problem solving, watching an educational DVD, role play and taster sessions. This was followed by a one to one consultation with a nurse to classify their headache type and discuss medication, lifestyle factors and goal setting, and up to 8 weeks of telephone support.

The development of the education and self-management interventions is described in detail elsewhere [21].

We aimed to test the feasibility of the new intervention by running four groups each with up to 10 participants in community settings. We approached people who lived within easy travelling distance of proposed groups; participants provided written consent to attend the group intervention. We wanted to recruit and train two lay people and three nurses to deliver the intervention. The acceptability of the intervention was explored by conducting qualitative interviews with the participants who attended the groups and the facilitators that delivered the groups. Thematic analysis was used to identify common themes across the different components of the intervention.

We proposed that our primary outcome measure for the RCT would be a headache-specific outcome measure collected by postal questionnaire. We initially did a systematic review of the quality and acceptability of patient reported outcome measures for episodic and chronic headache disorders [9], and a qualitative review of the lived experience of chronic headache [22] to understand what outcomes are important to people with chronic headache. This process supported the short-listing of both headache-specific (Migraine-Specific Questionnaire v2.1(MSQv2.1) [23] and the Headache Impact Test 6-item (HIT-6) [24] and generic measures (EuroQoL EQ-5D-5 L) [25] and Short-Form 12-item Health Status questionnaire (SF-12) [26] to include in the feasibility study. However, the migraine-specificity of the MSQv2.1 [23] made it unsuitable for use with our chronic headache population. Therefore, with permission from the developers, the target attribute of ‘migraine’ was changed to ‘headache’ and the questionnaire renamed as the ‘Chronic Headache Quality of Life Questionnaire’ (CHQLQv1.0). We evaluated both the acceptability and psychometric performance (data quality, reliability, validity) of the modified measure against the HIT-6, EQ-5D-5 L and SF-12 [24‐26], providing the first evidence for the performance of the CHQLQ and HIT-6 [24] in a UK population and supporting selection for the RCT. Structured cognitive interviews were also conducted to explore the acceptability and relevance of the measures. Informed by good practice guidance, the interviews explored how responder’s made judgements when completing the PROMs, including aspects such as question comprehension, recall and ease of completion [27, 28]. The cognitive interviews and their analysis was carried out by an experienced qualitative team with expertise in this area.

We trained six research nurses to conduct the telephone classification interviews. Feedback from the training indicated that the nurses felt that the training workshop, opportunity to practice interviews and the training manual prepared them adequately to carry out the classification calls and that they gained confidence the more interviews they completed. Nurses and doctors from the NMC completed 111 and 108 headache classifications interviews respectively. There were 107 paired interviews. Median days between interviews was 32 (interquartile range, IQR, 21 to 48 days). Proportion of concordance of agreement between nurses’ and doctors’ interviews was 0.91, with moderate or very good agreement on PABAK agreement in main and sensitivity analyses respectively. Full details of these analyses are reported elsewhere [13].

We approached 85 participants to pilot the education and self-management programme; we were unable to contact 12 (14%) participants and 46 (54%) participants were unable to attend, reasons included work commitments, dates being unsuitable, home life (including childcare) and holidays (Fig. 3), 27 (32%) expressed interest in attending the intervention and of these 18 (21%) provided written consent to attend a group.

We piloted the CHESS intervention in four groups and with a total of 18 participants. The attendance at groups ranged from 3 to 6 participants and 17 participants attended the one-to-one consultation with the nurse. Qualitative interviews were completed with 12 participants using topic guides to explore participants’ experience of taking part in the intervention. On the whole the groups were considered acceptable and participants found the educational and self-management components useful and interesting and found the opportunity to meet with other people with chronic headache particularly helpful. Based on participant feedback we removed the half day follow-up session because participants found the time commitment too great and we included the sessions on communication and managing setbacks at the end of day two of the programme.

Facilitators gave us feedback in a focus group or interviews with the use of topic guides, including their experiences of delivering the intervention and the training received. They reported that they did not find the two- day training adequate time to cover the delivery of the group intervention and the headache classification and medication information for the one-to-one consultations. It was also difficult for the lay facilitators to commit to delivering the intervention due to existing work and family commitments and unpredictability of their own headaches.

Participants completed and returned 131 baseline questionnaires; 115 (88%) and 103 (79%) questionnaires were returned at two and 12-week follow up respectively. Measurement data quality, reliability and validity for the headache-specific and generic measures was reached at acceptable standards [29, 30], supporting application of the measures with groups of patients with chronic headache. Participants in the cognitive interviews (n = 14) indicated items included in the CHQLQ were comprehensive in scope and particularly welcomed those referring to the emotional impact of headache, and found the measure easy to complete. The lack of recall period for the first three items of the HIT-6 was a concern. The generic measures were considered to be acceptable.

In total eight participants downloaded the Smartphone App, participants completed the app for a duration of up to 11 weeks. A telephone call was made to a selection of participants to check they were happy using the app and although participants didn’t report difficulties downloading or using the app only one participant completed all 11 weeks of data collection and only four participants completed half or more of the weeks.