A systems approach to developing user requirements for increased pulmonary rehabilitation uptake by COPD patients

Introduction

Chronic Obstructive Pulmonary Disease (COPD) is a progressive lung disease characterised by breathlessness and persistent cough. Patients report high symptom burden with associated anxiety and depression, reduced quality of life and loss of independence and freedom in an ever-shrinking world^1,2,3. Pulmonary rehabilitation (PR) is a “…comprehensive intervention based on a thorough patient assessment followed by patient-tailored therapies that include, but are not limited to, exercise training, education, and behaviour change, designed to improve the physical and psychological condition of people with chronic respiratory disease and to promote the long-term adherence to health-enhancing behaviours.”⁴ It is a cost-effective and transformative treatment for COPD⁵. PR reduces hospital admissions and leads to clinically significant improvements in breathlessness, fitness and quality of life^6,7. Referrals are mainly from primary care or from the community for patients with stable COPD (66.8%); 5.2% of referrals follow hospital admission for an acute exacerbation of COPD⁸.

In England, around 31% of referred COPD patients do not take up their PR appointment, more so from deprived communities^9,10. Systematic reviews have found insufficient high-quality evidence to make clear recommendations for effective interventions to increase uptake^11,12. In a previous study, we identified barriers and enablers to PR uptake at several stages on the PR pathway within and between primary care, PR providers and patients. In patient encounters with primary care, individual barriers to uptake, such as lack of confidence to exercise or not believing that PR will help, were often not addressed effectively. Patients needed a better understanding of how PR could benefit them personally and better support around emotional or practical challenges they faced in attending PR. Meeting such needs was a challenge for many primary care clinicians. There was an appetite for interventions that could address these needs¹³.

Facilitating the uptake of PR from primary care while meeting the needs of a wide range of users is a system-wide challenge involving patients, carers, GPs, practice nurses, physiotherapists who deliver PR and healthcare commissioners. Meeting the challenge requires a systems approach to take into account the needs of different users and the constraints imposed by complex supporting systems¹⁴. There is evidence for the success of systems approaches in healthcare, but they are not well established in healthcare intervention design¹⁵. The Engineering Better Care (EBC) framework for developing and improving healthcare settings is built on systems engineering principles that emphasise a holistic approach to design and are translated for healthcare applications¹⁶. Coproduced by NHS clinicians, managers, policy makers, patients and systems engineers, its development was led by the Royal Academy of Engineering in partnership with the Royal College of Physicians and the Academy of Medical Sciences. EBC is rooted in an established engineering model for the design and delivery of safe and successful systems. EBC principles have been refined into a how-to guide – the Improving Improvement Toolkit (http://www.iitoolkit.com). The Toolkit has been used to apply a systems approach to the design of a range of healthcare delivery challenges across the NHS and policy design across multiple areas of the UK Government.

The aim of the current study was to use the EBC framework to develop a set of user requirements for an intervention that could be delivered in primary care to increase uptake to PR.

Methods

Study design

The Engineering Better Care (EBC) framework incorporates strands of activity. These are illustrated in the Hexagon Improvement Model (Fig. 1). Agree the Scope, at the centre of the hexagon, underpins six outer activities. The six outer activities begin with Understand the Context followed by Define the Problem and so on in a clockwise direction. As new data are generated by each activity, outputs from previous activities may be refined. Within each activity, specific questions help drive exploration. Figure 1 shows the questions roughly mapped to the activities but this mapping is not definitive. The Improving Improvement Toolkit provides tools and techniques to help answer the questions. The questions are optional and only need to be asked when helpful, hence we chose a subset.

A systems approach to developing user requirements for increased pulmonary rehabilitation uptake by COPD patients — **Fig. 1: The Hexagon Improvement Model.**

The current study took place at an early stage in the design process—the development of user requirements—and so focused on the first two activities (Understand the Context and Define the Problem) and on the questions: “Where is the system?” “‘What are the elements?” “‘Who are the stakeholders?” “‘Who will use the system?” “What affects the system?” “What are the needs?” “What does good look like?”

This work was part of a broader mixed methods package of work that received a favourable opinion from the Cambridge Central Research Ethics Committee (REC reference number 17/EE/0136), was granted approval by the Health Research Authority and was registered with the ISRCTN registry (trial ID: ISRCTN20669629, assignment date 20 March 2018, trial start date 1 April 2016)¹⁷.

Four steps were followed to develop the user requirements (Fig. 2).

**Fig. 2: The development of user requirements.**

Step 1: Map the system

The expert study team (respiratory clinicians, systems engineers and applied health researchers) developed a system map of a typical pathway from primary care to PR to identify associated stakeholders and users and the system within which the pathway sits. They defined which areas of the pathway were in-scope for the intervention, i.e. where an intervention to increase uptake could be delivered in a primary care setting. In order to map and understand the system environment they drew on barriers and enablers to uptake previously identified by the team¹³, questions in the Hexagon Improvement Model (‘Where is the system?’ ‘What are the elements?’ ‘Who are the stakeholders?’ ‘Who will use the system?’ ‘What affects the system?’) and tools from the Improving Improvement Toolkit (flowcharting, storyboarding and expert reviews).

Step 2: Define user needs

The EBC approach describes user needs as things that users believe they must have. We distinguish these from user requirements, which are things that the solution (i.e. the intervention) must have in order to satisfy the user needs.

To answer the question ‘What are the needs?’ the study team mapped the barriers and enablers generated by stakeholders in their previous work to the system map, identifying those that were in-scope for an intervention to address. This previous work included patients who had declined PR, those who had never been referred to PR and patients who had attended PR¹³. In-scope barriers and enablers were those that a) mapped to areas of the pathway that were in-scope for the intervention and b) related to generic rather than local factors (focusing on generic factors would facilitate intervention rollout). Consensus was achieved through deliberative discussion. A further deliberative process was used to theme the in-scope barriers and enablers in order to define the user needs.

Step 3: Formulate user requirements

User requirements address the question ‘What does good look like?’ The study team used a deliberative process to formulate actionable user requirements that would answer this question and satisfy the user needs. While it would have been possible to create many user requirements at a fine level of detail, we required a number that was manageable for stakeholders to review in a workshop. This therefore determined their level of detail. For each user requirement, the team developed an evidence-based supporting statement that drew on the underpinning data in Early et al¹³. Each statement described a) the user needs addressed by the requirement; b) examples of related problems identified in the data and c) examples of what ‘good’ would look like if the user requirement was satisfied. The wording was informed by guidance on how to create good requirements¹⁸. For example, they should be clear and concise and should not overly constrain the design by specifying a solution, rather they should express an intervention end-goal that can be evaluated.

Step 1 and Step 2 were completed over 12 hours in four research team meetings (each with 3-5 team members present) held between October and November 2018.

Step 4: Validate and prioritise user requirements

User requirements and supporting statements were presented to stakeholders in a one-day workshop in November 2018. Users and providers of the system (people living with COPD and their carers/families, practice nurses, GPs, PR providers) and other stakeholders (commissioners, NHS England) were eligible to take part. Purposive and snowball sampling was used to recruit participants with knowledge and/or experience of the PR pathway. All participants gave written consent.

The aim of the workshop was for stakeholders to understand the user requirements and for the team to assess the variation in opinions about whether each requirement was ‘Essential’ or ‘Desirable’. A ‘not required’ option was not included because the requirements were derived from research data that represented known aspects of the system.

The workshop progressed as follows:

1.

Throughout the workshop, we aimed to promote an inclusive culture through a welcoming and encouraging tone. All participants introduced themselves at the start. The centrality of the patient voice was reinforced by beginning the workshop with a patient presenting his own story about his experiences with COPD and PR. Following this, the researchers presented two anonymised patient stories based on accounts from their previous research – one of a patient who had declined PR and one of a patient who had not been referred to PR – and a short vignette of an anonymised PR provider perspective of the patient experience at PR¹³.
2.

The workshop process was explained to the participants and a practice exercise conducted with a non-clinical example.
3.

Participants were divided into three mixed groups of 5-6 comprising patients, carers, clinicians and other stakeholders. Each group sat at a separate table with a facilitator from the study team. Groups were given A4 paper copies of the user requirements and supporting statements (Supplementary Notes 1). Each person had two A5 sized voting cards, one labelled ‘Essential’ and one labelled ‘Desirable’.
4.

Each user requirement, with supporting statement, was presented in turn verbally to the whole workshop. After each requirement, participants voted in their groups on whether it was ‘Essential’ or ‘Desirable’ by each person simultaneously holding up one of their voting cards. The facilitator recorded the results by hand in a matrix. The group then discussed the requirement for three minutes. Facilitators ensured that the patient and carer participants were able to contribute their views by inviting contributions from any participants who did not freely give a view. Participants then voted a second time in the same manner, the facilitator again recorded the results by hand in the matrix. This activity was a hybrid between nominal group technique, in which individuals work independently in the presence of one another, and a focus group technique to encourage discussion^19,20,21 hence participants voted twice – once independently and once after group discussion.
5.

Facilitators aggregated the votes from all the tables to produce an aggregated score for each user requirement as follows:

$$[(3* {rm{TE}}1){mbox{-}}(2* {rm{TD}}1)]+[(6* {rm{TE}}2){mbox{-}}1* {rm{TD}}2left.right)]$$

TE1 stands for Total Essential votes at 1st voting, TD1 stands for Total Desirables at 1st voting, TE2 stands for Total Essentials at 2nd voting and TD2 stands for Total Desirables at 2nd voting. In both rounds, more weighting was given to Essentials versus Desirables, as reflected in the constants chosen (3, 2, 6 and 1). In the second round, even greater weighting was given to Essentials over Desirables. This recognised the importance of the discussions in between the two voting rounds. The constant values were adjusted until the graphs appeared meaningful when applied to dummy data, i.e. accentuated the differences in importance between the requirements. Following this adjustment, the aggregated score value of each requirement ranged from 0 to 180.
6.

Results were presented to the workshop for verbal confirmation and questions.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Results

System map

Figure 3 shows the system map, illustrating a typical pathway from primary care to PR. The pathway begins with the patient’s symptom experience and diagnosis, followed by planned and ad hoc encounters with primary care. When a PR referral is made, the PR service contacts the patient to make an assessment appointment. If the patient attends the assessment, they are invited to begin the programme or excluded as not eligible. Not all local pathways are identical but our system map incorporates key generic elements that are likely to be represented in some way in most pathways. The shaded area is in-scope for this study, i.e. the area where an intervention to increase uptake could be delivered in a primary care setting.

User needs

Of the barriers and enablers to uptake we previously identified, twelve barriers and eight enablers were identified as in-scope for the intervention (Table 1)¹³. Three categories of user needs were derived from these: 1) Patients and clinicians needed to know and understand what happened on a PR programme; 2) Patients and clinicians needed to understand the benefits of PR as a treatment and how an individual patient could personally benefit; 3) Patients needed to be reassured about anxieties or concerns and feel positive about attending PR.

Table 1 In scope barriers and enablers identified by Early et al.¹³ translated into user needs and user requirement statements.

Full size table

Barriers and enablers that were out of scope related to patient-specific circumstances, primary care organisational factors, practical issues outside of primary care control (e.g., transport, health service organisation and delivery) and lack of clinician training in specialist respiratory skills (Supplementary Notes 2).

User requirements

Eight actionable user requirements were formulated from the user needs:

Requirement 1: Help the patient understand what happens on a PR programme

Requirement 2: Help the patient to feel positive about attending a PR programme

Requirement 3: Help the patient to understand how they will benefit personally from PR

Requirement 4: Help the patient to feel reassured about any anxieties they have about PR

Requirement 5: Help the HCP to understand what happens on a PR programme

Requirement 6: Help the HCP to understand the benefits that patients gain from PR

Requirement 7: Help the HCP to feel positive about PR and value it as a treatment

Requirement 8: Help HCPs, patients and carers to feel they have had a positive conversation about PR

The evidence-based supporting statements for the requirements are shown in Supplementary Notes 1.

Validation and prioritisation of user requirements

Eighteen participants (15 female, 3 male) took part in a stakeholder workshop: 5 patients (4 female, 1 male), one carer (female), 6 physiotherapists from PR providers, 3 GPs, one practice nurse, one commissioner of PR services and one representative from the British Lung Foundation.

Figure 4 shows the number of votes cast for ‘Essential’ and ‘Desirable’ for each user requirement before and after group discussion. User requirement 1 (help the patient understand what happens on a PR programme) was voted ‘Essential’ before and after discussion by all participants. For other requirements, the number of votes for ‘Essential’ or ‘Desirable’ changed slightly before and after discussion. A breakdown of raw scores is shown in Supplementary Table 1 and Supplementary Table 2.

Figure 5 shows aggregated scores for each user requirement after the two rounds of voting. Specific values are shown in Supplementary Table 3. The numerical values from 0-180 are arbitrary and have no objective meaning; their value is to provide a consistent way of measuring participants’ responses. The user requirements were ranked in importance as follows:

**Fig. 5: User requirement aggregated scores.**