Background

The number of new cancer cases is estimated to have reached approximately 300,000 per year in the UK by 2020 [1]. As many of these cancers may be preventable through reductions in lifestyle risk factors such as smoking, obesity and alcohol consumption [2], health promotion is increasingly becoming the focus for policy and clinical interventions. Alongside strategies to engage clinicians and commissioners, interventions to increase public awareness of cancer risk factors and promote behaviour change are required.

In the UK, primary healthcare already delivers the largest disease prevention initiative, the NHS Health Check programme, and also provides annual reviews for individuals with chronic diseases. Within the NHS Health Check programme, eligible individuals aged 40–74 are invited to attend a Health Check every 5 years. The Health Check itself consists of three components: risk assessment, risk communication and risk management [3]. Risk assessment tools are used to estimate the individual’s risk of developing cardiovascular disease (CVD) and diabetes. That assessment is then used to raise awareness of relevant risk factors and inform discussion on the lifestyle and medical approaches best suited to managing the individual’s disease risk. Chronic disease reviews for diabetes, asthma, hypertension and chronic obstructive pulmonary disease include assessment of the current condition, review of blood tests and current medication and discussion about associated lifestyle factors.

As many of the risk factors for the most common preventable cancers (such as lung, colorectal and breast cancer) are the same as those for CVD and chronic diseases, the NHS Health Check programme and chronic disease reviews provide an ideal opportunity to deliver interventions encouraging behaviour change for cancer prevention to large numbers of the population. The inclusion of cancer may also increase the impact of the NHS Health Check on CVD and diabetes: while studies of the impact of provision of CVD and diabetes risk suggest that risk information influences decisions around medication but not health-related behaviours [4], there is greater uncertainty surrounding the impact of cancer risk information [5]. Furthermore, recent studies have reported that both patients [6] and primary care clinicians [7] would welcome the inclusion of cancer into both NHS Health Checks and chronic disease reviews, with primary care clinicians reporting that discussions about cancer are part of their working roles and primary care is an appropriate setting for these.

Using behaviour change theory, reviews of existing literature and expert opinion, we have developed a brief intervention (the I-CaPP intervention) to facilitate discussions about cancer risk and prevention within primary care [8]. In focus group discussions and online usability testing with 65 healthcare professionals (HCPs) currently involved in prevention activities in primary care within one Clinical Commissioning Group in the East of England, the intervention prototype was described as potentially acceptable and one that would fit within current practice and may encourage patients to accept recommendations and motivate them to change their behaviour [8]. However, there is a need to assess the feasibility of its delivery in practice. In particular, concerns were expressed about the additional consultation time and resources required.

The aim of this pilot study was to assess the feasibility and acceptability, both to HCPs and patients, of implementing this brief intervention incorporating cancer risk information to promote behaviour change within NHS Health Checks and chronic disease reviews in primary care.

Methods

Study design

The study was a mixed-methods non-randomised pilot study [9]. To enable us to compare the recruitment of participants and feasibility of the delivery of the intervention in both NHS Health Checks and chronic disease reviews, we recruited patients into several study groups. Two groups received the standard NHS Health Check or chronic disease review and two groups received either an NHS Health Check or chronic disease review plus the I-CaPP intervention. In all groups, clinical measurements and lifestyle information were collected as in usual care.

Study patients and recruitment

In order to enable us to observe any differences between practices and have a sufficient pool of participants to recruit for the qualitative interview component of the study, we planned to recruit five general practices and 80 participants, 60 of whom would receive the intervention. The NIHR Clinical Research Network: Eastern supported recruitment of the five general practices. Three additional practices were also recruited but they withdrew before the start of patient recruitment due to other commitments. The profile of each practice was obtained from the Public Health England General Practice profile online database [10], and the decile of deprivation extracted based on the Indices of Deprivation document published by the UK government in September 2019 [11], with the lowest decile indicating the most deprived population. Additional File 1 shows the characteristics of each of the five practices. All practices were about the same size as or larger than the national average practice (8726 patients) [12]. The deprivation index for each practice ranged from six to 10, with four of the five practices in the two least deprived deciles. All served a predominantly white population.

Potentially eligible patients were identified from electronic searches already in use by the practices to identify patients eligible for NHS Health Checks or chronic disease reviews. These searches were conducted by each practice, with a GP screening potential patients to ensure eligibility prior to the invitation. Patients are eligible for an NHS Health Check if they are between the ages of 40–74, do not have a diagnosis of CVD, diabetes, kidney disease or hypertension, and have not had an NHS Health Check within the past 5 years. Patients are eligible for a chronic disease review if they have been diagnosed with the relevant disease and have not had a review in the previous 12 months. Eligible patients for this study were those over 40 years of age who were due a NHS Health Check or annual chronic disease review, did not have a current diagnosis or medical history of cancer or dementia, were not known to suffer from psycho-social issues or severe illness, in the GP’s clinical opinion had a life expectancy of more than 1 year, and were able to provide written informed consent. Patients not meeting all these eligibility criteria were excluded from the study.

Eligible patients were invited into the study by invitation in the usual method by their general practice (Fig. 1). These varied by practice and included sending personalised letters via post with or without follow-up with a telephone call by practice administrative staff or text message reminder (Additional File 1 and Fig. 1). Study clinics were held on different days of the week and included the morning and afternoon sessions to facilitate attendance.

Fig. 1
figure 1

Recruitment flow chart

Full size image

I-CaPP intervention

The I-CaPP intervention is modelled on the three components of the NHS Health Check. The risk assessment and risk communication components comprise an online risk assessment tool that is used within the consultation to provide the patient with an estimated 10-year risk of developing one of the most common preventable cancers based on their current lifestyle [9]. The common preventable cancers are specific to each sex. For women they are breast, lung, colorectal, endometrial and kidney cancer, and for men they are lung, colorectal, bladder, kidney and oesophageal cancer. The risk management component includes guidelines for the HCPs to deliver lifestyle advice for risk modification, including the opportunity for the patient and HCP together to set targets and see a revised risk estimate. At the end of the consultation, the patient is then given a behaviour change leaflet, a copy of the risk assessment, and website log-in details to enable them to re-visit the website post-consultation.

Practice set-up and training

Practice set-up included an initial meeting between the research team and practice staff, including both clinical and administrative teams, and a 30-min face-to-face training session for all HCPs involved in the delivery of the intervention. Nine HCPs (4 practice nurses and 5 healthcare assistants) were trained to deliver the intervention. At four of the five practices two HCPs delivered the intervention, at one practice one HCP delivered the intervention. During the recruitment period two HCPs left their roles at the practice and so recruitment continued but with less consultation and resource availability at these practices. Training focused initially on risk factors for cancer, highlighting the overlap between these and CVD. It covered the rationale behind the inclusion of the intervention, to share cancer risk and promote behaviour change, and how this could be included within existing consultations. Guidance was then given on the use of the intervention, including practising with test patients to allow for familiarisation with the website and behaviour change leaflet. Written step-by-step guidance was also provided as a reference (Additional File 2).

Data collection

We collected data from four sources:

1. Audio/video recordings of patient consultations

With consent from both patients and HCPs, the study consultations were video or audio recorded to capture the delivery of the intervention.

2. Patient and HCP interviews

Patients who had received, and HCPs who had delivered, the intervention were invited to participate in semi-structured audio-recorded interviews. During the patients’ interviews, the discussion explored the impact of the intervention, particularly on perceived risk, response-efficacy and intentions to make behaviour change, alongside thoughts on the format, content and delivery of the intervention within the consultation (Additional File 3). The interviews with HCPs focused on the format, content and delivery of the intervention, the training received and the potential barriers and facilitators to implementation (Additional File 4).

3. Patient questionnaires

Patients completed three online questionnaires; at baseline, immediately post-consultation and at 3-month follow up. The questionnaires were specific to each patient group and included information on demographics, lifestyle risk factors and family history of CVD and cancer as well as validated measures of numeracy [13], time orientation [14], self-rated general health, CVD risk perception, cancer risk perception [15], cancer risk awareness [16], cancer-related worry [17, 18], anxiety [19], maladaptive coping [20], self-efficacy [21] and response efficacy [22] (Additional Files 5, 6, 7). Index of multiple deprivation (IMD) was derived from each participant’s home postcode and grouped into quintiles using 2019 English indices of deprivation data [11].

4. Process data from the intervention website

We collected quantitative data from the website on the time spent completing the questionnaires, the interaction with the website during the intervention, and whether patients returned to the web-based information after the consultation.

Consent

Patient consent

Patients provided online consent to complete the baseline questionnaire prior to the consultation. Written consent was then sought from each patient immediately before the consultation. This included consent for the audio or video recording of the consultation and the post-consultation and 3-month follow-up questionnaires. For the patients who agreed to have their consultation recorded, a confirmation of this consent was also taken on completion of the consultation. Patients who took part in qualitative interviews completed written consent prior to commencement of the interview.

HCP consent

Written informed consent was obtained from each HCP at the start of the study. At the end of each recruitment session, the HCPs provided additional written consent for each of the recorded consultations. The HCPs also provided written consent prior to completion of qualitative interviews at the end of the recruitment.

Analysis

Audio/video recordings of patient consultations

The audio and video recordings were used to assess the fidelity of intervention delivery and the time taken to deliver the intervention. A fidelity checklist was devised based on the main elements of the intervention covered in the training session for HCPs. Using the video or audio recordings, two researchers (KM and BP) piloted the checklist with four consultations. One researcher (BP) then assessed the remaining consultations. Both mandatory and optional elements of the intervention delivery were assessed and subsequently summarised descriptively.

Patient and HCP interviews

Patient and HCP interviews were analysed using thematic analysis [23]. Using an iterative analytic approach from the beginning of data collection, each interview transcript was first repeatedly read by one researcher (KM, a research associate with qualitative expertise) in order to identify patterns within the data and develop a coding framework. The qualitative dataset was then fine-coded within NVivo software (QSR International, version 12) by the same researcher and the codes sorted and combined to generate themes. These initial themes were then reviewed and refined alongside two further researchers (JUS and FMW, both academic GPs with qualitative expertise) who had each read a selection of the interview transcripts.

Patient questionnaires

Descriptive statistics were used to summarise the baseline characteristics of the patients from the self-completion questionnaires and the follow-up rates. The potential effects of the NHS Health Check/chronic disease review alone or in combination with the I-CaPP intervention were then summarised by reporting the difference in mean values (follow-up minus baseline) with 95% confidence intervals for each continuous outcome and the proportion in each category at each time point for categorical outcomes. The views of the participants who received the intervention on the lifestyle advice and risk information provided were also summarised. As this was a feasibility study, no formal statistical tests were performed.

Process data from the intervention website

The data from the website on time participants spent completing the questionnaires, the pages viewed during the consultation and whether participants returned to the site after the consultation were summarised by reporting the mean with 95% confidence intervals.

Results

Feasibility of recruitment

In total, 62 patients were recruited between June 2018 and March 2019. The response rates for each patient group are shown in Fig. 1. Thirty-four patients (5% of those invited) attended for an NHS Health Check plus the I-CaPP intervention; 7 (6% of those invited) attended for a standard NHS Health Check; 16 (25% of those invited) attended a chronic disease review plus the I-CaPP intervention; 5 (10% of those invited) a standard chronic disease review. The characteristics of the patients in each group are detailed in Table 1. The mean age of the patients was comparable between the patient groups (56.2 years for two groups receiving the I-CaPP intervention and 55.0 years for the NHS health check/chronic disease review standard consultation groups). Fifty-two percent and 72% respectively were female, over 90% were White British, and over 88% had completed at least secondary education.

Table 1 Baseline characteristics of patients
Full size table

One patient withdrew from the study during the study consultation, and another withdrew prior to completion of the 3-month follow-up. Three did not consent to audio or video recording. Seventy-percent (n = 44/62) of patients completed the immediate follow-up questionnaire and 60% (n = 37/62) the questionnaire at 3-months. There were no marked differences between the characteristics of those who completed and did not complete the questionnaires at each time point of the study.

Feasibility of delivering the intervention

The mean duration of the consultations in each group are given in Table 2. There was variation between practices in the time taken for both the standard consultations and the time to deliver the intervention. The mean time taken for delivery of the I-CaPP intervention within the NHS Health Check was 9.6 min (SD 3, range 3.1–15.1) and within the chronic disease reviews was 9 min (SD 4, range 4.5–15.5). As a consequence of this time taken for delivery of the intervention, the mean duration of the NHS Health Check plus I-CaPP intervention was 23.6 min (SD 7, range 14.1–39.6), and 32.4 min (SD 14, range 11.1–62.8) in the chronic disease review plus I-CaPP intervention consultations.

Table 2 Length of consultations and use of website within consultations
Full size table

Based on the 46 consultations that were audio or video recorded, overall the fidelity of delivery of the intervention was high (Fig. 2). HCPs verified patient responses to the risk factor questions prior to the risk calculation in 43/46 (94%) of consultations and, although amendments were made in 84% of consultations, these were small and less than one unit for each risk factor. The HCPs used a variety of descriptions of the risk presentation, with 89% using phrases in keeping with the intervention delivery training such as “This is based on people who are like you, this is not necessarily yourself – it’s people with your height, weight, ethnic category - those sort of things”. Initial goal setting was completed with 96% of patients, with the new risk being communicated to 76%. For those patients with risk scores in line with those following the lifestyle guidance, HCPs tended to encourage them to continue with their current lifestyle and so there was less requirement for completion of target setting. In some cases, the HCPs continued to complete this section suggesting small improvements to lifestyle to ensure that the risk level was optimised. Despite inclusion in the training, only 26% of consultations included a discussion of the overlap between the risk factors for cancer and cardiovascular disease. Optional features of the intervention that HCPs used included the display of percentage risk (59%), lifestyle advice accessible via the hyperlinks provided on the website (48%) and the option to write goals into the website (11%).

Fig. 2
figure 2

Fidelity of intervention delivery

Full size image

Acceptability of the intervention to HCPs

Seven HCPs (3 practice nurses and 4 healthcare assistants, with at least one from each included practice) participated in interviews. Six themes (in italics below) were identified (Table 3). In general, HCPs were positive about the overall delivery of the intervention and the specific components. Many were encouraged by the potential for the intervention to promote behaviour change with specific reference to the ease of communication of the risk presentation. Discussion also focused on the ease of patient engagement in a shared activity, and using the computer to view the potential effects of lifestyle changes on risk of cancer. All felt this to be the most effective element of the intervention. Alongside the patient’s current risk, demonstrating a change in risk visually on the screen enabled the HCPs to introduce specific behaviour changes and show their impact in real time. HCPs found sharing lifestyle information via the website and providing a written leaflet to be useful additional resources for patients to consider after the consultation.

Table 3 Illustrative quotes of healthcare professional views
Full size table

The addition of cancer risk information was described by some HCPs as having an impact on patients’ responses to risk information and willingness to consider lifestyle changes. Many felt the format of the risk presentation to be of huge value in communicating information about risk of cancer. Some highlighted how the colour coding displayed on the graph enhanced patient understanding. Including the percentage risk was felt to be less helpful as most HCPs described how the scores were interpreted by patients to indicate low risk and hence not something to be concerned about.

The intervention was reported by most to be suitable for inclusion in NHS Health Checks but there were mixed views on its integration within chronic disease reviews. To ensure effective implementation into normal practice within NHS Health Checks and chronic disease reviews, the HCPs considered factors that would be of benefit. These included integration in the practice computer software and a more holistic conversation around risk of diabetes, cancer and cardiovascular disease and associated behaviour change. Additional suggestions to refine the intervention training included greater provision of information on specific types of cancer.

Acceptability of the intervention to patients

Twelve patients (Table 1) participated in interviews. Six themes (in italics below) were identified (Table 4). Prior to the delivery of the intervention, some patients reported knowledge of the risk factors for cancer. They referred to specific risk factors that did not apply to them such as cigarette smoking, which they felt ensured that their risk was low for both CVD and cancer. Patients reflected on their expectations of their risk. Most felt that the risk level presented was as expected but seeing it made them focus on the impact of their lifestyle choices on their risk of cancer. The risk presentation was generally quite well understood and they felt that the format facilitated interpretation. All described how the colour coded graph on its own or alongside the display of the risk percentage was helpful. This was in contrast to the communication of CVD risk which was noted to only be delivered verbally as a percentage. The visual display of the risk modification was reported by some patients to act as a motivator for behaviour change. Among those patients, those with a relatively low risk reported feeling motivated to maintain their current lifestyle and those who needed to make behaviour change felt better informed and had been provided with suggestions on how to make small changes to reduce their future risk. The patients also expressed how the provision of cancer risk information within the NHS Health Check was acceptable to them and fitted well with CVD risk communication.

Table 4 Illustrative quotes of patient views
Full size table

Patients’ views on the risk information and lifestyle advice are summarised in Table 5. Overall, patients agreed or strongly agreed that the lifestyle advice presented in the intervention was understandable (85%), trustworthy, (76%), useful (82%), motivating (80%), important (72%) and well-presented (79%). The risk information was also felt to be understandable (94%) and trustworthy (89%). Four patients re-visited the website. All who re-visited entered additional targets into the website to view the impact of further modifications on their future risk of cancer. Two patients also viewed the lifestyle information webpages.

Table 5 Views of patients who received the intervention on the lifestyle advice and risk information
Full size table

Potential effects of the intervention

Immediately after the intervention, participants reported high levels of intention to change behaviour across all the risk factors discussed (Table 6). There were also several potential effects on lifestyle factors and modelled risk of cancer at 3 month follow-up (Table 7). In particular, amongst those who received the intervention, reported fruit consumption increased (mean change 0.67 portions per day, 95% CI 0.1 to 1.23), and there were decreases in red meat and processed meat consumption (mean change 0.47 portions per week, 95% CI − 0.87 to − 0.09; and − 0.35 portions per week 95% CI − 0.68 to - 0.02 respectively). There was also a reduction in estimated relative risk in comparison to the average person (mean change − 0.34, 95% CI − 0.56 to − 1.28). These changes were not observed in those who received the standard consultations without the I-CaPP intervention. These changes were accompanied by an increased awareness of cancer risk factors immediately following the consultation (mean change 2.9 95% CI 0.10 to 5.0) and 3 months post consultation (mean change 1.58, 95% CI 0.04 to 3.13). There was no evidence that the intervention increased levels of anxiety or worry or that it improved accuracy of risk perception (Additional File 8).

Table 6 Intention to change behaviour immediately post intervention delivery
Full size table
Table 7 Lifestyle and estimated risk-related outcomes
Full size table

Discussion

We have shown that incorporating a theory-based, risk communication-based brief intervention to promote behaviour change for cancer prevention into primary care consultations is feasible and acceptable to both patients and HCPs. In particular, patients receiving the intervention found the cancer risk information and lifestyle advice presented to them understandable, useful and motivating, and HCPs delivered the intervention with high fidelity and felt that it fitted well within their practice. Although this study was not designed to demonstrate efficacy, our findings also suggest that the intervention has the potential to produce behaviour change without increasing worry or anxiety. This pilot study, therefore, provides support for a future randomised controlled trial of an intervention incorporating risk-based cancer information and linked behaviour change advice into primary care.

However, there were a number of challenges that arose during the study and should be taken into account when interpreting the findings of the pilot study and designing future trials and other studies of behaviour change and cancer prevention in primary care. The biggest challenge was recruitment. From the 630 patients invited to receive the intervention within NHS Health Checks, only 34 patients (5.4%) attended. A further 96 (15.2%) attended for an NHS Health Check outside the study. These numbers are substantially lower than the 57.5% uptake of NHS Health Checks for the first three quarters of 2018–19 across our region of recruitment [24]. They are also lower than uptake in a trial of a physical activity intervention nested within NHS Health Checks in the same region in which 27% (n = 373/1380) attended an NHS Health Check outside the study and 14.1% (n = 194/1380) took part in the study [25]. There are several reasons why our recruitment may have been lower. Firstly, although it was optional, eligible participants may have been put off by the inclusion of video/audio recording of the consultations. This is supported by the similarly low uptake (6%) among participants invited only for a standard NHS Health Check without the cancer intervention. Secondly, in order to enable a researcher to be present to consent participants and set up the recording equipment and to allow for the additional consultation time required for those taking part in the study, clinics were pre-allocated within each practice. Although these were arranged on different days of the week and at different times, this lack of flexibility to arrange an appointment may have acted as a barrier to recruitment, particularly for participants who were employed or who had caring responsibilities [26]. We were unable to collect data on participants who would have liked to take part but were unable to get an appointment. As reported in other studies [25], there was also variation in uptake between practices. This may have been due in part to the different invitation methods used by the practices: telephone invitations to NHS Health Checks have been shown to be more effective in previous studies than letter invitations [27]. The response rate among participants invited to receive the intervention within chronic disease reviews was higher (25%, n = 16/65). This may be because these individuals are accustomed to being invited for annual reviews. Both of these potential explanations suggest that the low levels of recruitment reflected aspects of the research design rather the intervention itself. To address these challenges with recruitment, we suggest that the invitations and appointment booking for future interventions are integrated within existing practice systems and letter invitations enhanced with behavioural insights [27] where telephone invitations are not feasible. The need for additional data collection or recording should also be restricted to only that needed to directly assess implementation.

A second challenge was the time needed to deliver the intervention. This had been identified as a potential challenge by HCPs during the development of the intervention [8] and in previous studies exploring the potential for incorporation of cancer prevention within primary care [7, 28, 29] and health promotion activities more generally [30]. In designing the intervention we had, therefore, sought to minimise the contact time required. This included enabling patients to enter the risk factor information prior to the consultation, developing a website that would auto-populate with that information, and providing a leaflet and facility to print a summary that enabled patients to go back to the website after the consultation. These components were all felt by HCPs to be helpful but the mean time to deliver the intervention of 9 min was almost double our target of 5 min, with a range from 3.1 min to 15.5 min. Some of this time was spent checking and amending the information provided by participants prior to the consultation and moving between the electronic healthcare record and the study website. Integrating the risk assessment and communication components of the intervention into the electronic healthcare record may reduce this time and support HCPs [31] and so should therefore be a priority for any future studies.

A third challenge identified by the HCPs was how best to include the discussion about cancer risk and behaviour change within chronic disease reviews. The intervention used in this study had been modelled on the three components of the NHS Health Check (risk assessment, risk communication, risk management). HCPs felt it fitted well within that context but that a clearer introduction may be needed within chronic disease reviews where risk of disease is not discussed as explicitly. Integrating the intervention within the electronic health record may also help with this, particularly if it were incorporated within existing templates.

Despite these challenges, the views of both patients and HCPs on incorporating the intervention within primary care consultations were generally very positive. In particular, the HCPs involved in this study reaffirmed the findings of previous research that has shown how prevention activities are felt to be an important part of their role [7, 28, 29, 32]. The high fidelity of delivery of the intervention also reflects the high levels of engagement among the HCPs and the potential value they saw in the intervention. Unlike HCPs in other studies [28, 30, 33], the HCPs in this study did not appear concerned about a lack of efficacy of the intervention or discuss the need for long term follow-up and referral services. This may have been because the focus of this study was on their role and how the intervention integrated within their practice. Patients also reflected positively on both the delivery and content of the intervention. Few though went back to look at the website after the consultation, questioning the value of that element in future interventions.

These findings must be interpreted in the context of the limitations. A key limitation is the fact that the five general practices included were self-selected and all in one geographical region in the East of England, with four of the five practices in the two least deprived deciles, and all serving a predominantly white population. We cannot therefore comment on whether the challenges with recruitment or the views of HCPs and patients would be different in other regions and among areas with higher deprivation or different ethnic characteristics. The participants themselves were also mostly of low deprivation and well-educated and many were already following a number of the lifestyle recommendations. This mirrors findings for uptake of NHS Health Checks in general [34]. They had also self-selected to attend and so may have been more health aware or concerned about their risk of cancer than the general population and we only collected self-report data on health-related behaviours that could be influenced by social desirability bias. Understanding, if possible, why eligible participants chose not to take part in the study and collecting objective measures of behaviour would be of value in future studies.

Conclusion

This pilot study of a theory-based brief intervention shows that incorporating discussions about cancer risk and lifestyle advice to promote behaviour change for cancer prevention within primary care consultations is feasible and acceptable to both patients and HCPs. A randomised controlled trial is now needed to assess the effect on health behaviours. When designing that trial, and other prevention activities within primary care, there is a need to consider potential challenges around patient recruitment, the HCP contact time needed for delivery of any interventions, and how best to integrate discussions about cancer risk seamlessly within routine care.