A tailored psychological intervention for anxiety and depression management in people with chronic obstructive pulmonary disease: TANDEM RCT and process evaluation

Cognitive–behavioural therapy improves anxiety and depression

Cognitive–behavioural therapy has been widely evaluated in a broad range of mental health conditions, and it has been adapted and delivered face to face or using a variety of digital health options. (The Cochrane Library lists 168 reviews relating to CBT in anxiety, depression, severe mental illness, substance misuse, self-harming, obsessive compulsive disorders and post-traumatic stress, among many others.) Psychological therapy based on CBT principles for people with anxiety disorders is effective in reducing anxiety, worry and depression symptoms, at least in the short term. 30–34 Similarly, psychological therapy, including CBT, reduced depression scores over the medium/long term35 and, importantly, given the age group affected by COPD, was effective in older patients with depression. 36 On the basis of this evidence, clinical guidelines now recommend CBT (via the internet or face to face) for the treatment of mild to moderate depression25 and generalised anxiety. 26

Benefits of cognitive–behavioural therapy for people with anxiety and/or depression associated with long-term conditions

There is increased recognition of the importance of identifying and treating mental health problems in people with long-term physical conditions. 37,38 Policy initiatives, such as promoting Improving Access to Psychological Therapies (IAPT) services,39 build on systematic review evidence that CBT can reduce symptoms of depression in people with physical conditions (including COPD). 40 Patient selection is important as effects are greater in people with clinically relevant depression at baseline. 40,41 Benefits for symptoms of anxiety have also been demonstrated in people with cardiovascular disease42 and cancer,43 both of which are common comorbidities in people with COPD. A lower-intensity CBT intervention [termed a ‘cognitive–behavioural approach’ (CBA)] delivered by physiotherapists to groups of people with low back pain has also been shown to be effective. 44

A Cochrane review,45 with 13 studies and 1500 participants, and focused on people with COPD, concluded that psychological therapies (using a CBT-based approach) may be effective for treating COPD-related depression. 45 There was a small effect showing the effectiveness of psychological therapies in improving depressive symptoms when compared with no intervention [standardised mean difference (SMD) 0.19, 95% confidence interval (CI) 0.05 to 0.33; six studies, 764 participants] or with education alone (SMD 0.23, 95% CI 0.06 to 0.41; three studies, 507 participants).

Benefits of pulmonary rehabilitation

There is robust evidence from a Cochrane review,50 with 65 studies and 3822 participants, that completion of a course of PR significantly improves health-related outcomes, including increased functional exercise capacity, reduced breathlessness and improved quality of life, and this specifically includes an improvement in the Chronic Respiratory Questionnaire score for ‘emotional functioning’ (mean difference 0.56, 95% CI 0.34 to 0.78), suggesting benefit for people with both COPD and anxiety/depression. These effects are described as ‘moderately large and clinically significant’. 50

In addition, people with COPD report benefits from attending PR beyond health outcomes, such as having a better understanding of COPD and strategies that can help them live better with COPD, including improvements in social functioning. 51,52 Patients who had experienced these benefits wished that they had had been offered PR early in the course of their disease. 51

Barriers to attending pulmonary rehabilitation

There are well-documented delays in healthcare professionals initiating a referral,53–55 and more than one-third of patients referred do not attend the offered assessment. Likewise, assessment completion rates are persistently low,11,55 especially in people from the most deprived quintile, people who are underweight or very severely obese and people with more severe disease. 56 In addition to limited information about PR, and long delays between referral and starting a course, there are multiple practical barriers (e.g. travel, parking arrangements, timing of classes) that reduce people’s capability to attend and/or complete a course of PR. 57 People who had ‘lost hope’ or whose anxiety about their breathlessness ‘made exercise impossible’ were less likely to attend PR,53 suggesting that a focus on addressing these perceptual barriers could benefit people with COPD and anxiety/depression. Providing information and reassurance, addressing practicalities and maintaining contact while waiting to start the course are suggested as strategies to improve attendance at PR. 53

Strategies to increase uptake and completion of pulmonary rehabilitation

A systematic review of interventions (14 studies) to improve uptake and completion of PR had inconsistent outcomes,58 although uptake was improved when the referral was part of a care plan actively supported by the healthcare team59 or when information about the benefits of attending the course was provided to the patient. 60 Qualitative interviews reinforce the potential benefits of ‘repeated discussions’ to facilitate understanding of PR and its perceived benefits and of addressing any emotional and practical limitations associated with attending. 53 A systematic review of qualitative and quantitative evidence (48 studies) in relation to referral, uptake, attendance and/or completion in PR mapped interventions to aspects of the theoretical domains framework. 61 The domains most mapped were ‘environmental, context and resources’, ‘knowledge’ and ‘beliefs about consequences’, again, emphasising the need to support patients’ understanding of the benefits of PR. Potential practical solutions were also addressed (including provision of transport or parking, offering choice of timing of classes and language provision), along with organisational support to increase service capacity.

How patients and the public were recruited

Patients with COPD and carer advisors were identified via established patient networks. We had worked with some patients/carers before, whereas other patients/carers were introduced to us by study team members as patients who had previously expressed interest in involvement. In addition, we promoted the TANDEM study on social media, specifically Twitter, now renamed X (URL: www.twitter.com; Twitter, Inc., San Francisco, CA, USA). We approached potential advisors by letter, e-mail or telephone, or by attending one of their regular meetings and presenting the purpose of study, inviting them to become an advisor to the TANDEM study and giving options for involvement.

Most of the patients who expressed interest were members of support groups affiliated to the Asthma UK (London, UK) and British Lung Foundation (BLF) (London, UK) partnership (e.g. Breathe Easy patient groups in London, an exercise group in the West Midlands and a dedicated hospital-based PPI committee in Leicestershire). One patient responded to the Twitter invitation and we also recruited a patient representative known to a study team member through previous work.

Flexible options for involvement

We were flexible in our approach of involving patients as advisors. Advice was sought in groups (e.g. a discussion group to inform intervention development) or individually (e.g. to provide comments and feedback on intervention handouts). Communication could use any mode of contact (e.g. in person, telephone, e-mail, post).

Remuneration for involvement was discussed with Steven Towndrow (PPI/Engagement and Communications Officer, NIHR Collaborations for Leadership in Applied Health Research and Care North Thames, hosted by Barts Health NHS Trust). Vouchers to the value of £15 or £30 were provided based on involvement activity (i.e. £15 for reading and providing comments on study documents and £30 for attending a meeting).

The research cycle

The research cycle (Figure 3) highlights the contributions of PPI colleagues to the TANDEM study before, during and after the trial. A PPI co-applicant (who eventually withdrew from involvement in the study) had been with us from inception of the study. At grant application stage, the PPI co-applicant contributed to shaping the research question, methods and intervention, and helped write the lay summary. During the trial, the PPI co-applicant attended and contributed to the study team meetings and also attended the Independent Trial Steering Committee.

FIGURE 3.

The contributions of the PPI colleagues throughout the TANDEM study. Reproduced with permission from Kelly et al. 65 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original text.

Copyright © 2024 Sohanpal et al.

The nature of involvement from other patient or carer advisors varied, as it depended on the requirements at that stage of the trial. Types of involvement included the following:

• Providing comments and feedback on patient and carer study documentation (i.e. invitation letter, participant information sheet, consent form, study leaflets, interview topic guide, intervention handouts). Advice was sought on clarity, how easy it was to understand what is proposed and suitability of content and images, as well as suggestions for improvement.

• Providing comments and feedback on terminology, advising on the best words to describe the intervention.

• Providing comments and feedback on patient and carer recruitment and participant flow into the study.

• Providing comments and feedback on intervention content and materials, and intervention delivery processes.

• Role-playing with researchers to improve the process of approaching patients about the study.

• Completing the patient and carer questionnaire to assess burden and time of completion.

• Suggesting preferences for:

  • promotional study materials for study participants

  • an electronic database based on looking at screenshots of two types of electronic database.

• Commenting on dissemination plans.

The fit of the TANDEM intervention with the 2021 Medical Research Council guidance for complex interventions

The extended MRC guidance for development and evaluation of complex interventions published in 2021 continues to specify the four key phases: (1) intervention development, (2) feasibility testing, (3) evaluation and (4) implementation. 70 In addition, there are six core elements that should be considered within each phase:

1. How does the intervention interact with its context?

2. What is the underpinning programme theory?

3. How can diverse stakeholder perspectives be included in the research?

4. What are the key uncertainties?

5. How can the intervention be refined?

6. What are the comparative resource and outcome consequences of the intervention?

Although the TANDEM intervention was developed prior to this publication,70 each of these questions was addressed during the course of the developmental work. With regard to how the intervention interacts with its context, consideration was given to where the intervention would be delivered (e.g. within a patient’s home or at a clinic). The programme theory was extensively developed and illustrated in a logic model. Using the person-based approach meant that a full range of stakeholder perspectives were used to inform intervention development from the outset. The pre-pilot identified important uncertainties (e.g. whether or not acceptability and feasibility were influenced by qualification level of the facilitator) and the intervention was refined post pre-pilot. Finally, the TANDEM trial addresses questions of clinical effectiveness and cost effectiveness to understand resource implications.

Cognitive–behavioural therapy: a theoretical approach to influencing mood

The brief specified an intervention be designed that combining psychological therapy with physical retraining that was tailored both to the severity of the patient’s COPD (ranging from moderate to severe COPD) and their mental health (ranging from mild to moderate anxiety and/or depression)’. 62 Arguably, the most commonly used and well-proven psychological therapy for people with mood disorders, either with or without physical health conditions, is CBT25,26,38,74 (see Chapter 1 for an overview of CBT). CBT typically addresses unhelpful cognitive interpretations that negatively influence mood and feelings. By working at the levels of thoughts, behaviours or symptoms CBT aims to interrupt and reverse the negative impacts and, hence, improve mood. A CBA has been effective when delivered by healthcare professionals with expertise in the clinical condition and who are trained to provide low-intensity CBT (i.e. focused on dealing with current negative automatic thoughts, rather than ingrained historical thinking patterns29).

The TANDEM intervention and a cognitive–behavioural approach

The CBT model has, therefore, been shown to have relevance75 and has had some efficacy with patients with COPD,45,52,76,77 where interpretations of physical symptoms, such as breathlessness, and consequent coping (e.g. fear avoidance) are key precipitators of anxiety and depression. CBT was, therefore, selected as the psychological therapy of choice in the TANDEM intervention (see Chapter 1). Although the underlying theory was based on CBT, the skill set of the facilitators as experienced respiratory practitioners trained to deliver a brief course of treatment meant that high-intensity CBT intended to be delivered by a mental health practitioner would not be appropriate. 29 Instead, we considered that facilitators would use a CBA, working at the level of lower-intensity CBT and addressing primarily negative automatic thoughts, but not addressing engrained thinking patterns, such as rules or schema.

Linked pulmonary rehabilitation: additional benefits for anxiety and/or depression

Physical retraining delivered through PR is recommended for people with COPD. 2,3,46,78 PR has a strong evidence base and is known to improve both physical and psychosocial outcomes at all levels of COPD severity. 50 Therefore, it seemed appropriate to link the cognitive–behavioural element of the TANDEM intervention with existing PR services. The core aim of PR is to increase exercise so that Bandura’s self-efficacy model,75 which holds that confidence in ability to change behaviour is an important predictor of whether or not behaviour will be undertaken, is a useful theoretical basis. Bandura states that self-efficacy can be enhanced through mastery (i.e. successful performance of a goal), vicarious learning (i.e. seeing similar others perform the behaviour of interest) and persuasion (i.e. encouragement from a significant other). Bandura’s self-efficacy model was the theoretical basis of SPACE for COPD73 and so that alignment with the SPACE for COPD programme was appropriate. The TANDEM intervention, therefore, provided all intervention participants with SPACE for COPD materials.

The TANDEM intervention and supporting pulmonary rehabilitation

A challenge of PR is poor uptake and retention54,55,79,80 (see Chapter 1 for a summary of the main issues). The TANDEM intervention aimed to address the issues of poor uptake and retention by applying behaviour change theory to the intervention. We selected Leventhal and Leventhal’s81 common sense model of illness self-regulation, as this model is designed to help understand the dynamic processes that an individual goes through when understanding and managing illness. Leventhal and Leventhal81 hypothesised that to understand their condition, individuals develop cognitive illness representations, which are beliefs that revolve around the following six conceptual areas: (1) identity (i.e. the label and symptoms that are attributed to the condition), (2) consequences (i.e. the impact the condition will have on physical and psychosocial areas of the person’s life), (3) causes (i.e. what contributed to causing the condition), (4) timeline (i.e. how long the illness is perceived to last), (5) control (i.e. the extent the individual can exert control over the condition) and (6) coherence (i.e. the extent to which the condition makes sense). People also have emotional representations that, together with the illness representations, influence coping procedures. Coping procedures can include behaviours undertaken to manage a condition (e.g. COPD), such as ‘take medications’ or ‘attend PR’, as well as adopting problem-based coping strategies rather than avoidance strategies. Within the common-sense model, the impact of the coping procedures on both illness and emotional outcomes is then appraised with potential to feedback and influence early aspects of the model. Illness representations have been shown to be important in COPD,82,83 and interventions targeting illness representations have been shown to be successful;84 therefore, this theory was seen as relevant for the TANDEM intervention. A recent extension to the model is the inclusion of treatment representations (i.e. beliefs around effectiveness, worries and concerns of the recommended treatment). 85 Given that PR is a specific treatment that we were targeting, we felt it appropriate to include content around the theoretical construct of treatment representations in addition to illness representations.

The theoretical basis of the TANDEM intervention was, therefore, complex because of integration of several models; however, the models were considered to be complementary and such approaches have previously been used. 86

Recruitment

A maximum variation sample of respiratory healthcare professionals who had an interest in delivery of psychological interventions to patients were recruited through social media and professional networks. The healthcare professionals were invited to participate in either an individual interview (i.e. face to face or telephone) or a focus group, dependent on participant preference. Similarly, two focus groups for people with COPD and carers were arranged (one for individuals with experience of psychological therapies and one for patients without this experience). Patients and carers were recruited for the focus groups via a respiratory support group (‘Breath Easy’). These formal qualitative approaches were in addition to PPI input on, for example, the one-to-one approach, number of sessions and content of intervention booklets (see Chapter 1, Patient and public involvement in the TANDEM intervention, for further description of the whole integrated PPI contribution).

Topic guides and data collection

Interviews/focus groups were conducted by Liz Steed, Ratna Sohanpal and Karen Heslop-Marshall. Topic guides for both patient and professional groups included (1) perspectives on patients’ difficulties with living with COPD, (2) opinions on the proposed intervention (i.e. we described our ideas as outlined in the grant application to participants) and (3) key strategies that would maximise the chance of implementation if the intervention was shown to be successful. Interviews and focus groups were audio-recorded, transcribed and analysed thematically.

Participant characteristics

One focus group with respiratory professionals was conducted, comprising a respiratory consultant, three physiotherapists, an occupational therapist and one exercise practitioner. All participants in the focus group were employees at a tertiary care hospital. In addition, seven individual interviews were conducted with healthcare professionals [one general practitioner (GP), four psychologists and two physiotherapists]. These individuals all had experience of working with respiratory patients within the community, primary care or secondary care.

We were able to conduct only one focus group with patient and carers because of time and governance delays at the site where individuals who had previously received CBT were to be recruited. In total, four individuals with COPD, two individuals with other respiratory conditions and two carers participated.

Guiding principles

Table 1 provides a summary of the guiding principles for both the intervention and its implementation.

In general, both patients and healthcare professionals were very positive about the proposed intervention. Patients and healthcare professionals felt that anxiety and depression were important and relevant for individuals with COPD, and that anxiety and depression were closely linked to breathlessness, which might be an acceptable way to introduce the intervention to potential participants. There was recognition that different people would have different needs and, therefore, being able to offer an intervention tailored to the individual with the possibility of home-based sessions would be ideal. There was consensus that, given the physical complexity of these patients, delivery by respiratory professionals would be preferable to psychologists, but training and supervision would be essential. One of the key concerns was around the symptom of breathlessness and its causation (i.e. was breathlessness anxiety related or an indicator of a developing exacerbation?). For anxiety-related breathlessness, some basic distraction and breathing techniques may be appropriate; however, for an acute exacerbation, assessment and treatment by a healthcare professional would be needed. These management options are diametrically opposed. From a safety perspective, it was considered that training respiratory professionals (familiar with COPD and managing acute exacerbations) in psychological techniques would be preferable to training psychology practitioners in respiratory medicine.

Three-day training

For the TANDEM intervention, a 3-day training programme was proposed. The first 2 days of the training programme were delivered consecutively and were designed to cover background theory, introduction to depression and anxiety, and the basic knowledge and skills to use in the TANDEM CBA for patients with COPD. The training was designed to be delivered in a group by two trainers (a respiratory healthcare professional and either a health or clinical psychologist). Learning was interactive, with practical exercises and opportunities to practise skills with a professional actor who had been briefed on the training aims. This type of learning was informed by medical education curricula that suggests that providing students practice with simulated patients is beneficial to clinical competency. 92

A third day was scheduled 4–6 weeks later. During the interim period, all trainees were asked to apply the skills they had learnt within their clinical practice. On the third training day, trainees presented a case study of a patient to illustrate how they had conducted a formulation and used the skills they had learnt. The case studies presentations were discussed, with any difficulties or problems resolved.

Training manual

A comprehensive training manual was developed, comprising three sections:

1. A detailed outline of content and delivery recommendations for each topic.

2. A guide to skills of cognitive–behavioural and self-management assessment and intervention.

3. Background information on COPD and theoretical underpinnings of behaviour change and CBA for trainees who wished to gain additional knowledge.

All resources, such as patient handouts, were appended at the end of the manual. The aim was to make the manual a comprehensive ‘go-to’ resource for delivery of all TANDEM intervention elements.

Study materials (research and clinical documentation)

Facilitators were provided with all materials required to audio-record the intervention sessions, along with video instructions on how to (1) use the encrypted audio-recorder and (2) securely transfer audio-recordings to the study team. Research documents included a CBA-session contact and delivery log. Clinical documents (not for research purposes) were as follows:

• A CBA-session case notes and supervision document was used to record detailed clinical information (including personal identifiable details of participant) necessary for intervention delivery for each participant, and to facilitate clinical supervision. The clinical notes were stored securely and separately from the TANDEM research data.

• A summary letter that was sent to patients’ healthcare professionals (with the participant’s permission) following completion of intervention delivery. The letter included a brief summary about the intervention delivery, the PHQ-9 and GAD-7 scores (used by the facilitator to assess the mental health status of the patient at the beginning and end of the CBA sessions) and any other information that was important for healthcare professionals to know about the patient in relation to their ongoing care.

Trial training for facilitators and practical considerations

The facilitators were good clinical practice-trained to deliver the research aspects of intervention delivery, and a NHS letter of access was arranged if facilitators were seeing patients outside their usual trust. The facilitators were informed about the concept of blinding and the arrangements in place to prevent the researcher who would be collecting outcome data knowing the allocation of the patient.

The trial manager (RS) delivered training in practical aspects of intervention delivery, provided TANDEM intervention materials and organised brief telephone catch-ups following first patient assignment, and then approximately fortnightly, to resolve any queries or concerns related intervention delivery.

Beck Depression Inventory II and Beck Anxiety Inventory

The Beck Depression Inventory II (BDI II)110 and Beck Anxiety Inventory (BAI)111 are both widely used in psychiatric research. BDI II scores range from 0 to 63, with higher scores indicating worse symptoms. BDI II scores in the range of 14–19 indicate mild depression and scores in the range of 20–28 suggest moderate depression. 112 BAI scores range from 0 to 63, with higher scores indicating worse symptoms. BAI scores of ≤ 21 suggest low anxiety and scores in the range of 21–35 suggest moderate anxiety. 113

Health-related quality of life

Health-related quality of life was collected with the validated, respiratory-specific St George’s Respiratory Questionnaire (SGRQ). 114 The SGRQ measures health impairment in people with COPD, incorporating three subscales (i.e. symptoms, activity and impacts). 115 A total score is computed (out of 100), which represents the percentage of overall impairment (where 100 represents the worst overall possible health status).

Social engagement and support

To assess social engagement and support we used the five-question Social Integration and Support subscale of the Health Education Impact Questionnaire (heiQ). 116 Social activity was measured using an adapted version of the UK Time Use Survey,117 which listed activities people might do in their spare time (including social visits) and asked how many times, and for how long, the interviewee had engaged in the activity in the previous week and whether this was alone or with another person. 118

Brief Illness Perception Questionnaire

To understanding how the intervention might be working we used the Brief Illness Perception Questionnaire (B-IPQ). The B-IPQ is a nine-item questionnaire that examines the cognitive and emotional representations an individual holds of their illness. 119

Smoking status and tobacco consumption

Past and current smoking status and tobacco consumption and e-cigarette use was assessed using the Office for National Statistics survey report on smoking-related behaviour and attitudes,120 and the National Asthma and COPD audit programme. 121

Health status

The EQ-5D-5L was used to assess health status for the health economic evaluation. 122

Anomaly in allocation ratio

During the trial, an anomaly in the allocation ratio produced by the online randomisation system became apparent. Specifically, over the first 69 randomisations the observed allocation ratio was around 1.25 : 1, as expected, but over the next 70 randomisations the observed allocation ratio was around 5 : 1. The deviation from the expected allocation ratio during this second period of randomisations triggered a corrective action and prevention plan overseen by the sponsor and, with the approval of the Data Monitoring and Ethics Committee, that included migrating the randomisation system to a new platform. For the remainder of randomisations, an allocation ratio of 1 : 1 was specified to return the overall allocation ratio to a figure closer to 1.25 : 1, and the observed allocation ratio over this third randomisation period was close to 1 : 1. In addition, as an additional sensitivity analysis for the two primary outcomes, we included randomisation period (i.e. period 1, period 2 or period 3) as an additional covariate. We also present tables of baseline characteristics of participants in different randomisation periods (and in different trial arms) to confirm the balance of characteristics and to allow an exploration of whether or not these characteristics might have differed in different periods; however, as the trial recruited fairly rapidly, we did not expect this to be the case.

COVID-19 and impact of lockdown

In late March 2020, the UK Government imposed ‘lockdown’ restrictions to reduce the spread of COVID-19. We performed additional sensitivity analyses for the two primary outcomes to explore the impact of the COVID-19 pandemic on our findings. For this purpose, the data were split in two different ways: (1) according to whether data were collected before or after 19 March 2020 and (2) according to whether the mode of intervention delivery was fully face to face or fully/partially via remote sessions (or, alternatively, no CBA at all, in which case we could not define the mode of delivery). For each way of splitting the data, we conducted a sensitivity analysis that included the split as an additional covariate and we also investigated the split as a potential treatment effect modifier. In addition, we present tables of baseline characteristics of participants divided according to the same data splits to allow an exploration of whether or not characteristics of trial participants changed.

Training providers: competency assessment

On completion of the 3-day training programme, all facilitators underwent a role-play assessment with a professional actor (note that the same actor presented the same prespecified case to all facilitators). Trainees were given the task of conducting a 20-minute assessment and formulation with presentation back to the ‘patient’. All assessments were video-recorded and independently coded by two psychologists using the Cognitive First Aid Rating Scale (CFARS). 103 The CFARS is a validated measure for assessment of therapeutic competency based on the revised Cognitive Therapy Rating Scale. The CFARS includes 10 items, each with a 0–6 scale (with a maximum score of 60). Therapeutic competency is deemed to be achieved if a minimum total score of 30 is acquired with no individual item falling below 2. In the post-training assessment, we omitted item 9, relating to delivery of appropriate intervention, as we did not require facilitators to show application of change techniques within the role-play assessment and, therefore, set a minimum score of 27 to indicate therapeutic competency.

Treatment delivery

A bespoke fidelity treatment delivery framework was created for this trial, including the CFARS to assess therapeutic competency (minimum level rated at 30 because of the inclusion of all 10 items) and an intervention-specific adherence measure. The adherence measure included assessment of whether or not core components were delivered across each session, according to a 1–3 scale (1 = not at all, 2 = partial, 3 = complete), and, additionally, whether or not topic-specific sessions were delivered (0 = no, 1 = yes).

Up to two entire TANDEM therapy courses (i.e. all of the therapy sessions provided for one participant) were assessed for each facilitator. A random selection of courses were selected by random number sequence from the second to the fifth participants for each facilitator (the first participant was used when a facilitator saw only one participant). For facilitators who delivered sessions to more than nine participants, a random selection from the 10th to the 15th participants were selected. This assessment was used to capture therapy sessions delivered when TANDEM facilitators were less experienced and when they were (relatively) more experienced. For each participant, all sessions within the course were coded. In the event of missing/unclear evidence from the audio-recording (e.g. when it was unclear which resource was being discussed), then this could be clarified from the facilitator-completed trial case report forms that noted, for example, provision of leaflets.

Coding was conducted by a psychologist (VW) who was trained in the psychological intervention and cognitive–behavioural methods more generally. Seven (19.4%) cases were duplicate coded by a member of the study team (LS) for quality assurance; however, to ensure consistency, the scores of the primary coder were used for analysis.

Treatment receipt and enactment

Qualitative interviews were coded for treatment receipt. A quantitative measure of social engagement (i.e. the heiQ)116 was used as a proxy indicator of treatment enactment. Enactment will be further explored in the qualitative interviews with patients.

Content of sessions and practicalities

Information and self-management skills provided in the TANDEM sessions increased awareness and better understanding of COPD and anxiety/depression. The practical tools were helpful for reflection and application in managing health conditions. Resources, such as leaflets and handouts, were sometimes considered too much, but could be referred to if needed.

The ‘hot cross bun’ formulation diagram was typically introduced in session 2, although, on some occasions, it could not be introduced until as late as the fourth session. In addition, the diagram took some time to complete, as patients needed to be guided to identify and explore their thoughts, feelings and behaviours in a new way. This was also true of homework tasks. Exploring emotions and behaviours was not always straightforward, therefore, the efforts of facilitators to build a therapeutic alliance as the basis for delving into new areas with patients was important to help patients to engage in this process. The format of the sessions was considered acceptable by the patients:

I remember the doughnut [the TANDEM hot cross bun], about how to try and go about things to make sure that I was less out of breath and options like stopping ... she also ... I had, which I’ve still got, two videos talking through different experiences and what would I do if say I was stuck in here and I wanted to go down to the bottom of the garden, and I couldn’t make it, ways to get over that.

Female, 66–75 years, anxious and depressed, moderate COPD, PAT14

Many of the interviews were conducted during the COVID-19 pandemic. Some participants were coping reasonably well, whereas others felt that the pandemic was having a negative effect on their health. The TANDEM model (and content) was new to participants and was not felt to be comparable to other NHS services currently available to them. Specifically, the option of having sessions in their homes, and discussions on mood and COPD, including practical advice and management strategies, were highly valued by participants, and participants felt that the TANDEM intervention could make a useful addition to their care.

The carers’ role comprised provision of personal care, practical and emotional support in line with the needs of the participant, and this affected carers in various ways. Opinions were divided about whether or not the CBA sessions were beneficial for the participant:

No, not at all, no. No changes really [after the TANDEM sessions for patient participant].

Male, spouse, 66–75 years, CAR2

Experiences of the TANDEM intervention

Patients developed a good rapport and a positive relationship with their facilitator over the course of the CBA sessions. Several participants seemed surprised about how much the facilitator was able to get them to open up and talk about issues that they had not considered or did not want to talk about previously. The same view was also expressed by a carer. The TANDEM intervention made some patients realise the connection between physical and mental health for the first time and gave patients confidence to move forward and to see the value of life again:

Other things she [the facilitator] reinforced was ‘don’t over think things, but analyse your reaction or your feelings, and try and figure out why you react a certain way’. And then if you consider that it’s not the way you want to act, you can find an alternative strategy to accept what’s happening.

Male, 76–85 years, depressed, moderate COPD, PAT16

But he seemed to open up, they seemed to have a nice chat in there. She was here about an hour I think, a good hour. Perhaps a bit longer some days. I don’t know, I can’t remember now. But it was just the regular activity and he used to sit in the room on his own and I used to leave him to it, because I knew he’d open up more.

Female, 66–75 years, spouse, CAR4

A few participants said that they were not comfortable reading handouts and/or writing things in between sessions for home practice and, therefore, the facilitator had modified their approach, which was helpful.

Two interviewees completed no sessions. For two interviewees, the sessions stopped at the right time because they were not affected by their mood, and were already familiar with the information provided on COPD. For one patient, the participant wanted to prioritise attending PR, as they had been unable to complete it previously because of a bad back and the programme sessions were going to overlap with the TANDEM sessions:

I sort of knew most of what she was saying really. I’d sort of been doing it myself. That’s the main thing. Went on the breathing thing [PR programme] for 6 weeks, and that was very good ... Yeah, it was sort of a joint decision [to stop the TANDEM sessions]. If I had any problems I could give her a ring. But then I more or less carried on with the breathing thing [PR programme].

Interviewer: And how soon did you join that exercise class after your one-to-one sessions?

More or less straightaway ... I think the thing for me was probably the actual breathing session ... Exercise and information.

Male, age unknown, anxious, severe COPD, PAT11

Therefore, five interviewees had completed fewer than four CBA sessions.

The effect of the TANDEM intervention on patients

The perceived effects of the CBA sessions for patients were wide-ranging. Participants described improvements in mental health, breathlessness and symptom management, knowledge, confidence and social life/activities, and in being able to accept their condition. Increased awareness about their own mental health was highlighted, as some participants realised or acknowledged for the first time that they had depression or anxiety. The TANDEM sessions were described as helpful to shift negative thoughts to positive ones and to improve feelings of anxiety relating to their COPD:

I didn’t realise I was depressed at the time because I was going to the bed most days in the afternoon, and that was more because I was ... I retired last year, so I think it’s because I’d recently retired. I was just bored. So I was going to bed in the afternoons just to kill time. And then I realised through talking to [facilitator’s name] that that was more me being depressed ... I’m going to say it changed my outlook on life. I didn’t realise how depressed I was and how down I was.

Male, age group unknown, anxious, moderate COPD, PAT6

And that’s [TANDEM sessions] helped a lot. So yeah, I don’t really get that anxious anymore ... I mean I’m not saying I don’t get anxious at all. I do occasionally. But not to the extent whereas before I’d be up all night.

Female, age group unknown, anxious and depressed, moderate COPD, PAT12

Participants also described learning skills/techniques and behaviours that some applied to improve the daily management of their COPD, for example integrating relaxation and breathing techniques to manage COPD-related anxiety. The sustainability of perceived improvements varied. The learning became embedded for some participants, whereas for other participants the positive effect was not sustained because of personal and external factors.

A few of the participants interviewed did not perceive any benefit. A couple of participants stated that nothing had changed or helped their condition and one participant did not believe that they could be helped and discontinued the TANDEM sessions:

I wouldn’t say it’s [the sessions] helped me, because I don’t think I can be helped ... to get better. Because I know I ain’t going to. But I live with it and put up with it. I’ve had cancer three times, so that ain’t beat me ... I didn’t expect to learn ... What can you learn about something you’ve got which ain’t going to have a cure?

Male, 66–71 years, depressed, moderate COPD, PAT17

Eleven interviewees completed the PR course and described health benefits. The four participants who did not attend/complete PR explained that this was because of competing comorbidities that needed to prioritised or because of health service-related barriers. Participants remembered discussions with the facilitator about PR focusing on how it may benefit them and what they could expect to gain by attending the programme. In a few cases, the facilitator helped participants to gain a referral to PR by liaising with participants’ clinical teams:

She [the facilitator] told me what it was all about and it was exercising to try and improve my breathing. And I did keep up the exercises for probably, I don’t know, 6 weeks maybe after the sessions finished. I don’t feel any worse for not doing them now. But at the time, again, it was something different. It was an hour or 2 hours a week out the house. We had information sessions and we had exercise sessions. And a little bit of competitiveness between the fitter versions of human beings there. So yeah, it was pretty good and I bought myself some dumbbells so I can exercise whenever I want to. I use the stairs. And step exercises on the bottom stair. If I go and use the loo, if I remember I’ll go up to the top of the stairs, turn around and come back down again and then go back up to use the loo. So I get double stairs exercise.

Male, 76–85 years, depressed, moderate COPD, PAT16

I think through [facilitator’s name] she also put me then in touch with the COPD exercise programme which I attended. So that was good.

Female, 66–75 years, anxious and depressed, severe COPD, PAT10

One carer felt that the patient had not applied any of the TANDEM learning recently because of the onset of the pandemic. However, the patient had attended PR after completing the sessions, which their partner found very helpful.

Interviewer: Before the lockdown, after the one-to-one sessions, was there a referral for [name] to go to the pulmonary rehab classes? The exercise classes?

Female, 66–75 years, spouse, CAR3: Oh yes, he did. He went to ... now let me get this ... You did, didn’t you? You went to the rehab clinic at [name] Hospital. Yes, he did, and that was very good, that was ...

The views and experiences of the telephone support sessions were explored among participants who completed PR; however, by the time participants were interviewed, the majority of participants were not able to remember this aspect of the support provided.

Complexity of patients’ lives

The fluctuation of COPD symptoms and multimorbidity had a cumulative affect on the physical, psychological and social health of the individual. Participants were aware that their COPD was not always visible to others, except when they suffered from noticeable symptoms, which led to feelings of being judged or embarrassment. Participants described social responsibilities that they prioritised over their own health, for example events that required them to look after loved ones and deal with traumatic events, which meant thinking less about themselves and their needs:

The thing that gets me down is I’ve got Parkinson’s as well, which is making walking difficult at the moment. It’s got much worse recently. And I get out of breath ... Although I’m walking slowly it takes a lot of effort.

Male, 66–75 years, depressed, moderate COPD, PAT5

I’ve got an awful lot ... still have, an awful lot of unfortunate stuff going on with my daughter ... But it’s not been a good time ... I know that’s not anything to do with COPD, but in effect it probably is because it’s all the stress and everything.

Female, 66–75 years, anxious and depressed, moderate COPD, PAT14

Carers reflected this complexity and described the care they provided as diverse, including a mix of personal care and practical and emotional support. The type of care they provided was dependent on what their partner might need for management of their condition(s) or and any changes in personality attributed to the partner living with a long-term condition, and this, in turn, could affect their relationship:

Well, yes. I think it’s more frustrating now for the both of us. Really it is frustrating because I suppose at the end of the day you don’t want ... I don’t want him to have dementia. So I’m angry. And I’m sure [name] is feeling depressed with the fact that he’s been diagnosed with the early stages of dementia.

Female, 66–75 years, spouse, CAR3

Effects of COVID-19

People living with COPD described a need to tap into their resilience and adapt to the lockdown restrictions, but overall they felt that their motivation was low. Ongoing or planned physical and social activities had stopped because of the pandemic and although some participants were drawing on techniques learnt in the TANDEM intervention to cope with their mood, other participants highlighted the negative impact on their health:

... I’ve brought into play some of my COPD ... the TANDEM thing. That way you look at your depression, why you’re depressed, let’s get out of it, let’s do something. So I actually wrote down things to do today and a timetable of what to do and when to do it. So I’m feeling a little bit better today ... I’m going to do things to make me feel better.

Male, unknown age group, anxious, moderate COPD, PAT6

A common concern of the carers (all interviewed during the COVID-19 pandemic) was the health of their partner in lockdown. Most daily and social activities had stopped, for example exercise stopped because of loss of interest or the inability to replicate the exercises learnt in their small space at home.

Experience of delivering sessions

Facilitators explained that it was important to build rapport and trust with patients in early sessions, developing a therapeutic alliance, which encouraged open communication in subsequent sessions. Although this was challenging, the TANDEM structure provided the necessary time and space for this to happen:

So he wasn’t actually doing many activities to get breathless because he was avoiding getting breathless ... So while all we’d really done was talk about it and allow him to open up about being scared about being breathless, by him talking about it changed his whole perception of it, and he realised that he could do more. Without even setting small goals he kind of just moved on in leaps and bounds, and was getting on public transport and everything.

Nurse 1

Patients’ engagement with the TANDEM intervention was variable, and facilitators mentioned factors, such as illness perception, age and illness trajectory, that could have contributed to this. Talking with the facilitator helped patients to make connections between their feelings and their COPD, although some patients did not feel that they were anxious or depressed or did not find their breathlessness an issue and may, therefore, have engaged less as a result:

It’s quite interesting to see that they don’t feel that the COPD is their main issue. They think that it’s the other stuff. But then actually when you’re talking to them then they realise that actually the COPD is a problem and that it’s been a bit of an eye opener for them.

Occupational therapist 1

Most facilitators highlighted how the complex lives of patients with COPD affected their day-to-day lives, including management of COPD. Social factors and multimorbidity were prominent in people’s lives. Facilitators worked with patients to identify and address concerns relevant to the patient in a case management approach within the context of the TANDEM intervention. A holistic patient-centred approach to psychological care was viewed as important in tackling the complex medical and social needs of patients with COPD and for improving quality of life:

So one of the challenges I’ve had is that a lot of my patients, breathlessness is a problem. But it’s not necessarily their main problem, because they’ve got lots of other things going on in their lives that have had huge impacts that you can’t really put totally to one side.

Physiotherapist 1

Contact with carers

Some carers (usually partners) sat in on TANDEM sessions. A few facilitators commented on their experience of carers being present during the TANDEM sessions. In some cases, involving the carer positively contributed towards the patient–carer relationship, whereby the facilitator was able to help the carer understand the patient’s experience of living with a chronic condition like COPD:

On the occasion where it was formally done I think it was useful because there was a definitive discrepancy between the thoughts and expectations of the patient compared to the spouse ... think that the participant was more confident than the spouse about their abilities. But there was a mid-point really between the two, because I think the participant was perhaps a little unrealistic. But the spouse was a little bit negative. So I think there was a common bit in between that we were able to reach.

Physiotherapist 4

In some cases, the presence of the carer reassured the patient and the carer could facilitate the session if the patient had problems with speech, hearing or memory:

But the second patient, the more complex one, his wife was always present. And he wanted her present because he has got a bit of a hearing problem as well. And he says that ‘sometimes she hears and remembers things that I don’t. So I always like her with me’. It wasn’t a problem. She didn’t interfere and keep talking or anything like that. It wasn’t a problem. In fact, I think it helped.

Nurse 3

However, the presence of carers could sometimes obstruct the communication between facilitator and patient, although facilitators were usually confident in dealing with this.

One facilitator observed that it was similar to taking on a social support role, particularly when dealing with problems within the patient–carer relationship, and this may highlight a potential need for some training in managing patient–carer dyads:

I find being in someone’s house, telling the husband and the wife how they ought to relate to each other after 45 years or 50 years of marriage is a very unusual ... it’s a very unusual dynamic. Sometimes it almost felt like couple’s therapy. Which I certainly didn’t feel equipped for.

Physiotherapist 3

Experience of facilitator training

The overall experience of the 3-day training for facilitators was positive. Although facilitators found the training challenging, with a lot of material covered, the training prepared the facilitators for the complexities of this specialised role and, in doing so, developed their skills in providing psychological care for patients using a collaborative decision-making approach:

It was, yeah. And I think it was great that they made it so difficult. Because I have had participants who have been as challenging as that. And so it did help to feel that I was on the right track with that role-play and that videotaping. And I could use what I’d learnt from that process.

Nurse 2

Facilitators valued the tools they were provided with during the training, describing the importance of the tools and the time for detailed exploration of anxiety and depression with patients. Facilitators suggested some improvements to the training, such as observing or speaking to experienced TANDEM facilitators, learning about different patient scenarios and some of the challenges to expect (as well as how to address these challenges), and local support and feedback when first delivering the intervention.

Experience of supervision

Facilitators found themselves tackling difficult conversations with patients, which could be overwhelming at times given the complexity of the lives of many of the patients they saw. Supervision was generally considered to be essential in reinforcing their new skills, developing confidence in delivering psychological care and in identifying how to proceed with more complex cases. Supervision was also considered important to debrief when providing psychological care:

... there was one, that tricky lady with the other mental health problems. I was quite worried about her in terms of actually, was my input beneficial, or were there higher levels of stuff going on that were more complicated? So, speaking to [supervisor] about that and helping me work out the best plan of what to do to safety net my patient, that was helpful.

Nurse 5

Impact on professional role

Facilitators felt that it was important to address psychological, as well as physical, aspects of the illness experience of people living with COPD, and were keen to develop skills to address psychological health. This new skill set involved an intentional behaviour change to that of the facilitator’s usual clinical practice. Some facilitators found this change a challenge to start with, with some facilitators experiencing an initial lack of confidence in confronting patients’ mental health issues. However, facilitators also felt that their new TANDEM skill set had enabled them to work collaboratively in a patient-centred way with other patients, improving their clinical practice outside the TANDEM setting:

... from a nursing perspective I can help them physically, I can help you with your emotional psychological needs to a certain limit, but prior to TANDEM I didn’t know those skills and I was passionate and still am passionate about COPD. So the patients for me always felt that I’m really interested and I had that feedback when I left, it was brilliant ... So I imagine those who are interested but not looking outside the box to try and do it that way, they wouldn’t even think to try to find the skills that TANDEM could offer them and ultimately just help the patients.

Nurse 6

Conversely, some facilitators described that they found it challenging to step out of their normal clinical roles while acting as a TANDEM facilitator, and it could sometimes be difficult not to simply provide patients with answers or medical advice linked to their professional role. This was considered to be a potential problem if the TANDEM intervention was to be delivered as part of the facilitator’s usual clinical practice role.

With the time spent with each patient over a number of sessions and in telephone support sessions afterwards, facilitators were able to recognise how patients benefited from the TANDEM intervention and this, in turn, created a learning feedback loop for the facilitator.

Perceived enablers of implementation

There was appreciation that for many people with COPD there were several barriers to engaging with healthcare services, including, for example, difficulty travelling to appointments. Attending separate appointments with different healthcare professionals was highlighted as particularly challenging and, therefore, the number of services to which people are exposed should be reduced. Stakeholders felt that a strength of TANDEM intervention was the provision of integrated physical and psychological care within one healthcare service:

Physical and mental health, they come together. They come as a package. So, I think we should treat it as a package.

Participant 4 respiratory physiotherapist

Stakeholders recognised that reaching people at home improved the opportunity for engagement and enabled services to understand wider social determinants, which ultimately could improve uptake of PR:

When people go out to the house you often get all the added benefit of that social interaction... this person’s really struggling at home, we need to get the social work involved and stuff.

Participant 6 respiratory consultant

Furthermore, home visits enable services to understand contextual issues, such as social well-being of the patients, that could affect uptake of PR and be communicated to other primary care services:

And also, you can always feed back to us if they’ve got any issues at home. Because we can’t visit everyone.

Participant 1 GP

Several stakeholders were confident that respiratory professionals could be trained to deliver the TANDEM intervention and it was considered that the existing relationship with patients related to their physical care was a good basis for engagement with a psychological intervention:

But there’s a lot of trust straight away with the respiratory practitioner and I think already, because it’s coming from physical health initially, they’re happy then to engage with the mental health. Rather than starting the other way around really.

Participant 14 PR clinical lead/physiotherapist

Perceived obstacles to implementation

Stakeholders appreciated that the TANDEM intervention was an additional service to existing provision and that the TANDEM intervention would need to be resourced properly; however, concerns were expressed about the availability of resources and competing commissioning demands. There was agreement that the TANDEM intervention must demonstrate cost effectiveness and that this would be demonstrated by reduced healthcare utilisation:

And they will have to fight against their highly specialised respiratory colleagues ... who are more interested in bright, shiny, sexy respiratory [services].

Participant 11 GP and CCG commissioner

Does it work? Is it cost effective?

Participant 3 respiratory consultant and CCG commissioner

The key metrics are going to be, does it reduce hospital admissions ... and does it reduce GP visits.

Participant 2 respiratory consultant and respiratory lead

Organisational change barriers were also raised and it was highlighted that successful implementation would require ongoing support and supervision to ensure adherence to the approach:

But this stuff takes practice, doesn’t it? And you have to keep repeating the things that you learnt in order to make them your default setting. And I think that when people go back into their working environments, there’s so much pressure to behave like they used to...

Participant 3 respiratory consultant and CCG commissioner

Characteristics of facilitators

The professional status of the 31 facilitators who delivered the intervention were specialist nurse (n = 9), physiotherapist (n = 13), occupational therapist (n = 6) and physiologist (n = 1). All 31 facilitators had a respiratory background.

Receipt of the TANDEM intervention

Twenty-two (9%) intervention arm participants did not receive any CBA sessions, a further 24 (10%) received just one session and 136 (56%) received six or more sessions (see Appendix 1, Table 23, for details). The number of TANDEM intervention sessions received by the 242 participants in the intervention arm of the study is shown in Figure 17.

FIGURE 17.

Number of intervention sessions received by TANDEM intervention participants (n = 242).

From the caseloads of the 28 facilitators who delivered the intervention in the main trial (i.e. after the conclusion of the internal pilot in which the intervention was delivered by three additional facilitators), 37 cases were randomly selected for inclusion. Twenty-eight cases were selected from the first five participants and nine cases were selected from the participants 10–15. Consent for audio-recording was not available from one participant and so data are reported for 36 cases delivered by 27 facilitators. The median number of audio sessions per case was 5 (IQR 4–6), providing a total of 180 audio-recorded TANDEM treatment sessions.

Cognitive First Aid Rating Scale

The mean therapeutic competence (i.e. CFARS) total score across facilitators when delivering the intervention in the main trial (Figure 18) was 35.3 (SD 7.2). Twenty eight (78%) of 36 cases achieved fidelity, with a mean score of 38.2 (SD 5.4), which is comparable to other trials (a mean score ≥ 30 reflects performance on all 10 items). 162,163

FIGURE 18.

Mean CFARS score for the 36 cases (27 facilitators).

Mean scores for individual items on the CFARS are reported in Figure 19. Higher competency was observed for focus and structure of the session, collaborative relationship and interpersonal effectiveness. In contrast, facilitators demonstrated lower competency for guided discovery and application of appropriate change techniques.

FIGURE 19.

Mean scores for individual items of the CFARS for the 36 cases (27 facilitators).

Adherence to TANDEM intervention core components

All core intervention components across sessions reached > 80% adherence, with the exception of agenda-setting (Figure 20), and this is considered high fidelity. 161

FIGURE 20.

Proportion of cases reaching fidelity for the TANDEM intervention core components.

Anxiety content was delivered to 25 (69%) of the 36 participants and depression content was delivered to 21 (58%) participants. Fourteen (39%) participants received both anxiety and depression content. Four cases received no anxiety or depression content, as treatment was discontinued because it was no longer required.

Handouts

Handouts were provided by facilitators for patients to use between sessions. All 36 patients received the ‘control your breathing’ handout and 34 patients received the ‘mood and COPD’ handout. The remaining handouts, including handouts for anxiety and depression, were tailored to content (see Figure 21 for frequency of use).

FIGURE 21.

Proportion of participants provided with the tailored session handouts during the course.

Pulmonary rehabilitation-specific intervention content

Twenty-three of 36 (64%) participants confirmed their intention to attend PR. Discussing what to expect from PR was delivered in 20 (87%) of these cases and eliciting attitudes to PR (i.e. thoughts, feelings and expectations about PR) was delivered in 22 (96%) cases. The photobook (i.e. a bespoke book of photographs to enable facilitators to familiarise participants with their local PR services and staff) was used in only five cases and problem-solving was delivered in 35% of cases (Figure 22).

FIGURE 22.

Proportion of patients provided with the specific support to prepare for PR.

Participating in the TANDEM therapy sessions

• The format and length of weekly sessions were acceptable, and receiving the intervention at home (or having a choice of location) was often valued, although some participants felt that it did not help them.

• Participants appreciated the knowledge and skills of the facilitators, and liked the facilitators’ friendly, empathetic and supportive nature, which enabled participants to feel safe discussing their thoughts and feelings in relation to their COPD, other conditions or social situations.

• Initial TANDEM sessions and homework tasks involved patients taking time to consider their thoughts, feelings and behaviours in a new way. A few participants were not comfortable with the reading and writing involved, but the facilitators were able to adapt to this. Homework ‘tasks’ were often not completed.

• Practical tools were helpful for reflection and were used in management of COPD (and other conditions). The linking of mood and COPD with practical strategies was seen as very beneficial and increased awareness and understanding of COPD and anxiety/depression.

Unexpected themes relating to the complexity of patient lives

• Lack of understanding about COPD and its effects by family and society leads to feelings of being judged and embarrassment.

• COPD and multimorbidity have a cumulative effect on physical, psychological and social health, although patients have adapted and become resilient to living and coping with their condition(s).

• Patients often prioritised competing family and social responsibilities over their own health and needs.

• Carers provided a mix of personal, practical and emotional support based on the changing needs of the patient.

Impact of living with COPD during the COVID-19 pandemic

• Ongoing or planned physical and social activities stopped because of the COVID-19 pandemic, with challenges for the caring role. Some patients drew on techniques learnt in the TANDEM intervention to cope with their mood, but other patients described a negative impact of the pandemic on their health.

Practicalities of facilitating the TANDEM therapy sessions

• The TANDEM intervention was viewed as an acceptable and effective way of delivering psychological care to patients with COPD, which was generally seen as bringing value to patients.

• Training and clinical supervision were integral to embedding new skills and building confidence to deliver the TANDEM intervention, and, more specifically, enabled development of valuable skills in psychological care using collaborative decision-making.

• Most facilitators were not confident to provide psychological care initially, but felt that their competence developed over a series of sessions, which enabled them to experience a feedback loop where they could see impact on the patient. Clinical supervision was necessary for delivering psychological care and embedding the new skills.

• Most facilitators were able to manage the workload effectively, although some facilitators found it difficult to balance time on the TANDEM intervention with their clinical roles.

• The TANDEM intervention provided the time and space to deliver valuable holistic care using skills that could potentially enhance interactions in the Facilitators’ routine care, but there was discussion about NHS resource constraints.

Unexpected themes arising

• The social complexity of patients’ lives was highlighted. Through the TANDEM process, facilitators were able to identify the concerns relevant to patients in the context of their lives, which raised social, as well as psychological, issues.

• Some facilitators highlighted the challenge of managing TANDEM therapy sessions that involved carers, and, although this was not necessarily negative, specific training in this would be helpful.

Patient experiences and perspectives regarding the intervention

• Participants were positive about facilitators’ knowledge of COPD and mental health, and valued the facilitators’ personalised communication and therapeutic skills.

• Participants felt that they had established a good rapport and positive relationship with facilitators, and were able to open up and talk freely about their mood, which raised awareness for some participants that they had anxiety/depression.

• Most participants reported improvements in mental health, breathlessness and social activities, although some participants did not perceive any benefit. Perceived improvements became embedded for some patients and helped with managing their COPD, whereas for other patients the positive impact was not sustained because of personal and external factors.

• The care received in the TANDEM intervention was different from current services offered to people with COPD.

Facilitator experiences and perspectives regarding the intervention

• Holistic care, integrating mental and physical health was viewed as important in caring for people living with COPD.

• Facilitators felt that it was important to address psychological, as well as physical, aspects of illness, and engaging with patients to address anxiety and depression using collaborative, personalised care delivery had the potential to tackle complex medical and social care needs.

Patient and carer perspectives on implementation

• Uptake of the TANDEM intervention may be influenced by perceived need owing to a patient’s condition or level of anxiety/depression and perceived individual benefits.

• Flexibility in the location for intervention delivery was highly valued.

• Patients could potentially be offered the TANDEM intervention at different times over the course of the illness trajectory.

Facilitator perspectives on implementation

• Resource and organisational constraints are likely to impede undertaking the TANDEM intervention as part of usual clinical roles.

• Wider changes are needed within the current COPD political/healthcare system context for successful TANDEM intervention implementation.

• Most stakeholders agreed that the TANDEM intervention would be best delivered by respiratory health professionals, but training and supervision would need to be replicated at local level.

Organisational stakeholder views on implementation

• A strength of the TANDEM intervention was the provision of integrated physical and psychological care within one healthcare service, but it was an additional service and would need to be resourced properly. Clinical effectiveness and cost effectiveness would need to be demonstrated.

• Reaching people at home improves the opportunity for engagement and ultimately could improve uptake of PR.

Fidelity

The TANDEM intervention was delivered with therapeutic competency and key tasks were delivered with fidelity and, therefore, the lack of effect is unlikely to be explained by a failure to deliver the intervention as planned.

Uptake and attrition

Eighty-one per cent of intervention participants received at least two CBA sessions (i.e. our predefined minimal clinically effective ‘dose’ of the intervention) and 56% received six or more sessions (a full course of the intervention was usually six to eight sessions tailored to the needs of the participant).

Attrition from the study was relatively low and overall 93% of participants were followed up at 6 months (i.e. the time frame for the primary outcome) and just 13% of participants were lost to follow-up at 12 months, which includes deaths, withdrawals and patients loss to follow-up. At both time points, attrition was slightly higher in the intervention arm than in the control arm (6-month follow-up: 90% vs. 97%; 12-month follow-up: 80% vs. 88%, respectively).

Perceptions of the TANDEM intervention

Respiratory health professionals recruited to train as TANDEM facilitators recognised the need for holistic care for patients with COPD and were keen to develop their knowledge and skills in addressing psychological health needs. Facilitators considered their training to be a challenging but effective way of developing skills in psychological care.

The TANDEM intervention was generally well received by study participants. Although some participants did not feel that the TANDEM intervention could offer them anything, many other participants felt that the TANDEM intervention clearly enabled engagement with issues arising in the contexts of their (often complex) lives and many instances of positive effects on quality of life were described. Developing a therapeutic alliance was considered necessary by patients, carers and facilitators, but it took time to build rapport, and the complexity of the therapeutic task was highlighted in the patient, carer and facilitator interviews. The relationship between patient and facilitator built through the TANDEM process enabled therapeutic collaborative work to be carried out, linking physical and mental health with social aspects of people’s lives. Patients and facilitators considered the home environment to be most appropriate for delivery of the TANDEM therapy sessions.

Potential for implementation

Although most patients and health professionals felt that TANDEM improved health care and should be available for people living with COPD, some of the patients and most of the health professionals considered that resource constraints would affect post-trial implementation. Health professionals felt that it would not be possible to deliver the TANDEM intervention as part of their usual clinical respiratory health role.

Recruitment, attrition and outcome measures

To the best of our knowledge, this is by far the largest randomised controlled trial of any psychological intervention for people with COPD published to date, and one of the very few randomised controlled trials to include only patients with symptoms of anxiety and depression. 45 Recruitment to the study was very close to that predicted in our original application (i.e. we estimated that we might need to approach 4720 potential participants with COPD). Most participants were recruited via mail outs from their GPs, with 2191 (49%) participants agreeing to be approached by the research team to learn more about the study, of whom nearly half 1062 (49%) were found to be potentially eligible. A total of 441 (42%) of the participants screened were found to be eligible, most of whom joined the study.

We had low attrition, unlike several of the studies in the recent Cochrane review of psychological interventions for people with COPD, which reported high attrition rates. 166–171

The TANDEM intervention did not succeed in improving mood in participants with COPD with symptoms of mild to moderate anxiety and/or depression, as assessed by the HADS instrument. The HADS instrument has been found to be relatively poor at discriminating between anxiety and depression in COPD, although overall it appears to be a good measure of emotional distress and a useful screening tool in medical populations. 172,173 The HADS is one of the most common screening tools for depression in COPD studies and appears to perform as well as the BDI II and the other commonly used screening tools. 104,174 In addition, we measured BAI and BDI II scores at baseline and follow-up, and the baseline results for these instruments suggest that the HADS instrument successfully identified people with symptoms of mild to moderate anxiety and depression in our study.

Impact on mood

The TANDEM intervention failed to both improve mood (i.e. the primary outcome) and increase uptake and completion of PR (i.e. a very important secondary outcome). We had anticipated that if the TANDEM intervention had succeeded in both these outcomes them there would be synergy in the effect on mood, as PR also has as a positive effect on mood, improving both anxiety and depression. 50,175

We are aware of only seven other published randomised controlled trials that have examined similar interventions in participants with COPD and anxiety and/or depressive symptoms at baseline. Five randomised controlled trials were delivered to individuals71,167,176–178,179 and randomised controlled trials two investigated forms of group CBT. 166,180

Lamers et al. 167 tested a ‘minimal psychological intervention’ based on the Chronic Disease Self-Management Programme,181 Goldberg’s reattribution model (which is also part of CBT)182 and other work by the group. Set in the Netherlands, 187 participants with COPD were randomised to intervention or usual care. The intervention was delivered by a trained nurse in the patient’s home over 3 months in up to 10 visits. Participants had been formally screened for depression before joining the study. Although the severity of their COPD was not recorded, baseline BDI II scores suggested that participants had, on average, mild depression. After 9 months, but not after 3 months, the BDI II was significantly lower (i.e. improved) in the intervention arm of the study, as were symptoms of anxiety and SGRQ total scores. However, the relevance of the of the study is hard to determine, as no primary outcome was stated and attrition was very high (40% in the intervention arm).

Doyle et al. 176 conducted a small randomised controlled trial (n = 110), comparing telephone-administered CBT and befriending (as an ‘active social control’) for people diagnosed with COPD and mild levels of anxiety and/or depression in Australia. Both CBT and befriending were delivered over the telephone across eight weekly sessions following an initial rapport-building session. Unlike the TANDEM intervention, CBT sessions were delivered by registered, or provisionally registered, psychologists experienced in delivering telephone CBT. As in the TANDEM evaluation, there were co-primary outcomes measuring anxiety (using the BAI) or depression (using the PHQ-9). The study was underpowered according to the authors’ power calculation and no significant between-group differences for anxiety or depression were seen.

In Hynninen et al.’s180 small, underpowered trial (n = 52), intervention arm participants attended 7-weekly 2-hour group CBT sessions in a hospital outpatient clinic in Norway. At 6 months’ follow-up, the decrease in anxiety (BAI) and depression (BDI II) was greater in the intervention arm than in the control arm (note that patients in the control arm received 2-weekly 5- to 10-minute telephone calls from the study team).

A randomised controlled trial (n = 238) by Kunik et al. 166 compared mostly male (> 95%) participants with COPD and anxiety or depressive symptoms (as determined by the BAI or BDI II, respectively) randomised to either eight 1-hour group sessions or eight 45-minute educational lectures, each with 15 minutes for discussion about COPD. Both the CBT and lectures were delivered by psychology interns and post-doctoral fellows with experience of delivering CBT. Attrition was very high (> 50% at 4 months) and there was no significant difference in anxiety, depression or quality of life (i.e. the primary outcome) at all follow-up periods up to 1 year.

Bove et al. 177,178 led a small randomised controlled trial (n = 66), comparing a 1-hour home visit psychoeducational session ‘based on... CBT’ and a telephone ‘booster session’ 2 week’ later, involving Socratic questions and discussion of models linking thoughts, emotions, sensations and behaviours, compared with usual care. Study participants had severe COPD and symptoms of anxiety (HADS-A scores ≥ 8). Not all baseline HADS-A scores were collected prior to randomisation. At 1 and 3 months post intervention, there were significantly better HADS-A scores in the intervention arm than in the control arm, but the risk of bias was high, as outcomes were collected in person by the principal investigator who also delivered the intervention.

Heslop-Marshal et al. 71 randomised 279 participants with COPD and symptoms of anxiety (HADS-A scores ≥ 8) to receive either an active control (a self-help leaflet addressing anxiety management) or a brief CBT intervention of 2- to 6-fortnightly sessions delivered in either outpatients or the patient’s home by a trained nurse. The primary outcome was change in HADS-A at 3 months’ follow-up. Attrition at 3 months was low and there was a statistically significant reduction in HADS-A score of a size suggesting clinical significance in the CBT arm compared with the leaflet arm at 3 months. At 6 and 12 months’ follow-up, the difference in favour of the CBT participants was much smaller and not clinically significant.

In a randomised controlled trial, Pumar et al. 179 attempted to evaluate CBT (first two sessions face to face followed by four sessions via telephone) delivered by psychologists to people with chronic lung disease undergoing PR who screened positive for anxiety and or depression. Approximately 75% of participants had COPD. Slower than anticipated recruitment led to the study falling short of its intended sample size, with 65 participants rather than 100 participants randomised in total, and only 24 participants were randomised to the intervention. Attrition was high, especially in the intervention arm and only nine intervention arm participants provided data at 12 months’ follow-up. No significant differences in anxiety or depression were seen between participants in the study arms at 3 and 12 months’ follow-up, but the small sample size makes the findings hard to interpret.

Overall, the Cochrane review of all psychological interventions for people with COPD and depressive symptoms concluded that there was some evidence for the effectiveness of psychological therapies with a CBT‐based approach in depression associated with COPD. However, the effect sizes were small and the evidence was of low quality and at risk of bias. 45 The review called for larger, more robust studies that consider AEs, health service use and cost-effectiveness outcomes. 45

Developing the intervention

A possible explanation for our results could lie in the content or focus of the intervention (i.e. if the intervention was not suited to the needs of people with COPD). Our intervention, however, was built on evidence-based and guideline-recommended CBT,25,26,30,45,76 adapted to a low-intensity approach,29 in line with effective CBA intervention in other clinical contexts,44 and linked with potentially synergistic evidence-based PR. 50 Given our extensive efforts at PPI and theoretically based development, this seems unlikely. Moreover, the content of the TANDEM intervention and the aims of our evaluation are strongly endorsed and in line with the research and healthcare priorities of patients with COPD identified in a recent Canadian survey. 185

Ensuring and measuring fidelity

Other potential explanations for our results could lie in the fidelity with which the intervention was implemented or its uptake within the trial. We undertook a comprehensive assessment of fidelity and are confident that fidelity was high and maintained throughout treatment delivery, indicating that recruitment and training of facilitators was effective and that the intervention was implemented as planned. However, our intervention was delivered by trained healthcare professionals, not trained psychologists and, although our measures of fidelity suggested that overall the facilitators delivered the intervention well, their lack of a background or previous experience in psychological interventions might have reduced effectiveness, even though they received regular supervision from experienced clinical psychologists. Perhaps the psychological complexities of some participants were beyond the scope of the facilitators? This is possible, although the Cochrane review of psychological therapies for COPD, which found some weak evidence that such interventions may improve depression in people with COPD, included both interventions delivered by psychologists and by trained nurses. 45 There is also evidence from other areas that healthcare professionals can deliver cognitive–behavioural interventions. For example, a systematic review44,186 of interventions for low back pain suggested that, with suitable training, physiotherapists could deliver effective cognitive–behavioural interventions.

An adequate dose of the intervention

Even a well-designed and well-delivered intervention may have a reduced affect on participants if they do not receive an adequate dose. Our predetermined minimal effective ‘dose’ of two sessions was based on the finding that, in some cases, improvement has been reported following two sessions of a psychological treatment,187 and 81% of participants achieved this, although, in retrospect, this may have been too low a threshold. Only 64% of participants received at least four sessions and 56% of participants received six or more sessions. In addition, it was noted that, although both facilitators and participants interviewed were positive about the intervention, there was some lack of alignment of facilitator and patient (and carer) priorities at the commencement of the TANDEM sessions. Agenda-setting was one area that was not successfully delivered, as observed in the fidelity analysis, although a ‘collaborative relationship’ was established overall. The health professionals (i.e. facilitators and stakeholders) commented on the unmet mental healthcare needs of patients with COPD, framing this as a care deficit. Most patients (and carers), on the other hand, did not identify themselves as being in need of psychological care. Patients were also frequently not used to talking about psychological issues and did not have experience of being offered psychological therapy, which contributed to the considerable work that needed to go into rapport-building in the early TANDEM sessions and making links between breathlessness and anxiety and/or depression. Producing a formulation often happened much later than anticipated in the course of sessions, which effectively reduced the sessions available to deliver the CBA intervention.

Therefore, it is possible that our intervention was not long enough or that a sufficient number of intervention participants did not receive a sufficient dose. None of the other published studies of interventions, which are similar to the TANDEM intervention, in participants with depression have reported how many session participants actually received the intervention and so it is difficult to compare our results with these studies. 70,166,176,180 In Heslop-Marshall et al.’s71 study of a nurse-delivered CBT intervention for patients with COPD and anxiety, the mean number of sessions delivered was four, which the author noted met the IAPT service specification for a brief intervention. 188 Heslop-Marshall et al. found the nurse-delivered CBT intervention to be effective in reducing anxiety, compared with a leaflet control, at 3 months. The process evaluation suggested that participants were not always keen on doing the home practice (the name TANDEM facilitators used to describe homework) associated with the CBA approach, which may have reduced the effectiveness of the intervention, as a meta-analysis189 has found that CBT is more effective with homework than without.

No impact on uptake of pulmonary rehabilitation

The TANDEM intervention failed to increase uptake of, or retention in, PR. Although disappointing, this is perhaps not surprising, as two recent systematic reviews58,190 of interventions to increase the uptake of PR revealed the paucity of evidence of effective interventions and called for further research in the area.

Complexity of lives and the challenge of multimorbidity

An important issue that may help to explain the trial outcome was the complexity of the lives of the patients, which emerged as a frequent theme in the qualitative interviews with patients and facilitators. This complexity included multimorbidity and social factors, such as caring responsibilities and previous trauma. As in previous studies,191,192 patients had lived with COPD for many years and had developed a level of resilience and coping strategies. Many patients did not necessarily consider their COPD to be their primary problem (or not at that point in time), nor believed that their anxiety or low mood was related to their breathlessness or COPD. In this complex context, the TANDEM intervention, which was delivered by respiratory healthcare professionals and designed primarily to address anxiety and/or depression related to COPD, may have had too narrow a focus. Certainly, many of the facilitators observed the breadth of the problems (relating to other conditions, historical issues or social context) presented by the participants.

We may have overestimated the power of the TANDEM intervention, that is a low-intensity CBA intervention delivered at a single, relatively late, point in a long-standing illness trajectory, when people’s problems were deeply embedded and when, over many years, they had established their own coping mechanisms to enable them to live with their slowly progressive respiratory disability. Significantly reducing levels of long-standing anxiety or depression in patients, like patients with complex multimorbidity, may have required a different or much more sustained intervention. It may be that there is a need for awareness-building about the psychological effects of COPD and the availability of support, such as CBA, to be woven into good long-term condition care from the time of diagnosis. 193 Although increasingly important in health care,194,195 complex multimorbidity involving physical and mental health conditions is particularly difficult to address successfully. 196,197

Adverse events

More AEs and SAEs were reported in the intervention arm of the study than in the control arm, but this is likely to be an artefact of the way that trial events are reported, as intervention participants had much more contact with the study through home visits and telephone calls with the TANDEM facilitators (who reported any AEs in the intervention group to the study team). In contrast, the healthcare resource use data (derived from primary care records) reported in Chapter 5 (see Table 16) showed similar or slightly lower mean inpatient stay bed-days, accident and emergency attendances and emergency admissions in the intervention arm participants, compared with the control arm participants. No AEs or SAEs were assessed as being related to the intervention.

Specifically, there were more deaths in the intervention arm, but none was assessed as being related to the intervention and the majority (9/13, 69%) arose in the second 6 months of follow-up (i.e. some time after the intervention was delivered). In addition, as we were able to obtain about only half of the medical records from general practices because of the pandemic, it is possible that some of the patients lost to follow-up had actually died. Owing to the contact the facilitators had with study participants, we would have been more likely to identify a death in the intervention arm than in the control arm.