A feasibility study of the CRISP intervention; a cardiovascular risk reduction intervention in patients with an abdominal aortic aneurysm

Patient and Public Involvement

This study was designed with the help of four patients with an AAA. Two lay individuals with an AAA were formal part of the research team and prepared/reviewed all study documents as well as this final report, including the plain English summary. A total of 101 patients have taken part in the qualitative work which led to the creation of the CRISP intervention. The Leicester AAA patient and public involvement group (eight participants) advised the research team throughout the conduct of this study.

Funding and approvals

This research was approved by the East Midlands Leicester Central Research Ethics Committee and the NHS Health Research Authority (HRA) and Health and Care Research Wales (HCRW) in January 2020 (reference: 19/EM/0366). The research was funded by the National Institute for Health and Care Research (NIHR) Academy (reference: NIHR300059) and sponsored by the University of Leicester (reference: 0746); the funder and sponsor had no input in data collection, analysis or interpretation. Participants provided written informed consent upon recruitment. This study was registered on ISRCTN (ISRCTN93993995; 18/11/2020) and supplementary material can be found within the trial registration page.

Intervention design and format

The CRISP intervention was designed following Medical Research Council (MRC) guidelines for complex interventions¹⁵, and using the Intervention Mapping framework, widely used in the development of health behaviour change interventions¹⁶. This involved a six-step ecological approach to assessing and intervening in this specific health-issue via the development of a new intervention and modification of existing interventions in similar areas, through a process of engaging patients and healthcare professionals to identify needs (behaviour change targets), barriers and enablers (determinants) of change and behaviour change techniques and strategies designed to modify the determinants identified¹⁶.

An overview of the CRISP intervention is presented in Figure 1 and a detailed description is provided elsewhere¹⁴. In brief; initially patients were asked to fill in questionnaires about their cardiovascular risk, and the following clinical measures were taken: blood test, height, weight and blood pressure. Responses were entered into a computer programme, which used the data to produce a personalised risk factor letter, given to the patient. The overall risk score was derived from the SMART risk score¹⁷ and the risk factors were graded as “red” (high), “amber” (medium) or “green” (low), based on risk-factor-specific algorithms. The patient then discussed their risk factor profile with a nurse in a “nurse assessment” appointment which occurred following AAA diagnosis. This was added to the existing initial nurse assessment component of standard care for AAA patients. In some programmes, the nurse assessment happens immediately following diagnosis, however, in others it occurs up to a few weeks post-diagnosis. The conversation during the nurse assessment covered the following: the patients understanding of and reactions to their risk factor profile; their motivations to modify their risk factor profile; which (if any) risk factors they would like to focus on. If the patient chose to focus on a risk factor(s) the healthcare professional would then discuss that risk factor specifically and give them the appropriate workbook(s). Then at every subsequent screening appointment, the healthcare professional would ask about how the patient is progressing with their chosen risk factor modification, problem-solve any challenges they presented, or discuss working on a new risk factor with the patient, or take no further action (if the patient is happy with their current risk-management plan).

Figure 1. Overview of the CRISP intervention.

Recruitment

Patients who were already part of an NHS AAA-surveillance programme were recruited to the study, across the East Midlands (UK), starting in September 2022 (01/09/2022), when all data collection began (quantitative and qualitative). The nurse assessment was then repeated with these participants to include the CRISP intervention components. They were then followed up at 6 months and the follow up part of the CRISP intervention was delivered. This is more frequent contact than normal for patients with a small AAA, who are currently seen yearly.

Inclusion and exclusion criteria

Inclusion: Adult males with an AAA (defined as maximal infrarenal aortic diameter above 3.0cm) detected via the existing NHS AAA screening programme, who are able to provide written informed consent. Exclusion: Cannot provide written informed consent; unable to comprehend written and spoken English; Body mass index (BMI) <18.5 kg/m².

Potential participants were identified by reviewing the list of men in aneurysm surveillance within the Leicester, Leicestershire and Rutland Abdominal Aortic Aneurysm Screening Programme (East Midlands). Potential participants were then approached, via letter, to take part.

Data collection

The following data was collected pre-intervention (baseline): age, gender, ethnicity, smoking (including e-cigarettes and vaping) status and history, healthcare setting of diagnosis, weight, height, blood pressure, medical and surgical history, eGFR, HDL, LDL, cholesterol, CRP, haemoglobin, list of medications, quality-of-life using the EQ5D5L, General Practice Physical Activity questionnaire (question 3 was not used), UK Diabetes and Diet Questionnaire, AUDIT C (alcohol screen), GAD-7 (stress), PHQ-8 (depression) tool. At six months and every time a patient came into contact with the screening programme the following additional data was collected: change in smoking, vaping, and alcohol consumption status, how many components of the intervention were received, date(s) of healthcare appointments. Data collection took place between September 2022 and May 2023.

Qualitative assessments

Semi-structured interviews were conducted with patients receiving the intervention who consented and with all intervention-provider staff. Topic guides were developed and piloted/refined across the first few interviews. Interview content explored participants’ and providers’ experiences of receiving and delivering the intervention, including barriers to engagement or delivery, acceptability and ideas for improving the intervention. Both patient and healthcare professional interviews stopped when data saturation was considered to have been reached. All intervention sessions were audio-recorded and audio files scored independently by a member of the intervention design-team (CG, TW) and an independent observer (AS), using an intervention fidelity (IF) checklist to assess quality of intervention delivery and the presence or absence of intervention components. The checklist used a Dreyfus competence-rating scale, providing a score from 0 to 5 for each of nine items which represented the key elements of the intervention’s content and the way the designers intended the intervention to be delivered¹⁸. The researchers scoring intervention fidelity took part in a calibration exercise, where they independently coded four initial assessment and four follow-up consultation sessions and then compared scores and notes.

Examples of good practice and areas for improvement were also identified (by time-stamp in the recording) and transcribed to inform future intervention training.

Analysis

Numerical data are presented as counts, means and standard deviations or median and range (where applicable). Interviews were analysed using framework analysis¹⁹, which uses both an inductive and deductive coding approach. The initial coding framework was based on the logic model of the CRISP intervention which is described elsewhere¹⁴.

Fidelity assessment

The fidelity of the study was generally low. The mean scores for all items in the fidelity checklist are summarised in Table 2. The highest mean score for the nurse assessment was 1.5 (SD 0.6) for engaging the participant, lowest 0.5 (0.4) for exploring the importance for selected lifestyle behaviours. For the follow-up appointments the highest score was 0.6 (0.3), reviewing progress with risk factors, and the lowest mean score was 0.0 (0.0) for exploring the importance for selected lifestyle behaviours.

When considering the fidelity scores by facilitators, a similar picture was observed, with little variation between facilitators at the nurse assessment (Table 3) and follow up (Table 4).

Inter-rater agreement was defined as a difference in fidelity scores between raters of one point or less and this was assessed for each checklist item for each recorded interview. A summary of inter-rater agreement for each checklist item is presented in Table 5. The sessions used for calibration (see Methods) were not included in this analysis.

Based on the quantitative and qualitative findings, the study team convened to collate a list of areas for improvement and change in order to improve the deliverability of the CRISP intervention for future use; the main issue identified by this group related to training of healthcare professionals delivering the CRISP intervention as well as central delivery of the intervention where possible, to streamline patient care.

Semi-structured interviews

A total of 25 (63%) participants and nine (100%) of healthcare professionals took part; this provided an ample size to draw conclusions from and data saturation was reached. There were three key themes: Benefits of the workbook approach, Personalisation of risk factor profile and Perceived risk factor reduction.

Theme one: Benefits of the workbook approach. The workbooks were perceived to be an efficient way to deliver cardiovascular risk reduction information as it balanced time efficiency with ensuring that the patients were sufficiently informed. The healthcare professionals felt the booklets were well written and easy to use.

The participants views were more mixed but still on balance positive. One participant noted that they “don’t really take much notice of books.’ However other participants viewed the workbooks as reaffirming that what they thought was healthy was correct, or as a useful tool for identifying small changes to improve lifestyle.

Theme two: Personalisation of risk factor profile. The personalisation of the risk factor profile was seen as beneficial and motivational to patients. Participants saw it as a moment of “realisation” of what they need to change or “affirmation” that what they already thought they needed to change to reduce their cardiovascular risk was correct. However, a number of participants did not remember the risk profile letter or workbooks they were given at the time of the interview. The healthcare professional perception of the risk factor personalisation was unanimously positive and there was a feeling that the information was represented in a relatable manner.

Theme three: Perceived risk factor reduction. Some participants felt that they had reduced their cardiovascular risk factors, including cutting down cigarette smoking, switching to vaping or losing weight. For example, one participant commented that “I’ve gone onto vaping.” Another stated “when I first met the nurse, I was 17 stone and now I’m 15 stone something” which represents a substantial weight loss over the 6-month study period. A separate participant noted that the CRISP intervention reminded them not to get complacent with regard to exercise.

However, one participant did comment ‘I don’t think it’s [the CRISP intervention] actually influencing what I actually do.’

Strengths and limitations

This study has a number of limitations that need to be noted. The study was conducted in a small, non-nationally representative sample (in terms of both patients and delivery sites). However, we recruited from both rural (Rutland, Leicestershire) and urban (inner Leicester) areas and incorporated in-depth quantitative and qualitative assessments. Due to funding constraints the study was delivered only to patients who could understand English. However, most men with an AAA are of white ethnic origin and the final format of the intervention will be available in a range of common languages.