Communicating cancer risk in the primary care consultation when using a cancer risk assessment tool: Qualitative study with service users and practitioners

Abstract Background Cancer risk assessment tools are designed to help detect cancer risk in symptomatic individuals presenting to primary care. An early detection of cancer risk could mean early referral for investigations, diagnosis and treatment, helping to address late diagnosis of cancer. It is not clear how best cancer risk may be communicated to patients when using a cancer risk assessment tool to assess their risk of developing cancer. Objective We aimed to explore the perspectives of service users and primary care practitioners on communicating cancer risk information to patients, when using QCancer, a cancer risk assessment tool. Design A qualitative study involving the use of individual interviews and focus groups. Setting and participants Conducted in primary care settings in Lincolnshire with a convenience sample of 36 participants (19 service users who were members of the public) and 17 primary care practitioners (general practitioners and practice nurses). Results Participants suggested ways to improve communication of cancer risk information: personalizing risk information; involving patients in use of the tool; sharing risk information openly; and providing sufficient time when using the tool during consultations. Conclusion Communication of cancer risk information is complex and difficult. We identified strategies for improving communication with patients involving cancer risk estimations in primary care consultations.


| BACKG ROU N D
Delayed diagnosis of cancer can adversely affect treatment options, outcomes such as survival and quality of life, and costs. [1][2][3] To tackle late diagnosis of cancer, there has been an increased interest in the use of cancer risk assessment tools to identify and quantify cancer risk in symptomatic individuals during primary care consultations. [4][5][6][7] Two cancer risk assessment tools designed for symptomatic individuals and available in primary care are QCancer [5][6][7] and the Risk Assessment Tool [RAT]. 4 Both QCancer and RAT were developed, tested, 4,5 and QCancer in particular has been independently validated, 8,9 as accurately quantifying risk of cancer in primary care.
Despite this, it is currently not clear how cancer risk information generated through use of the tools can be effectively communicated to patients.
Current evidence suggests that discussion of risk information was less likely to occur if patients did not prompt practitioners. 10 In a recent review of the literature around cancer risk assessment tools in primary care, a key challenge identified was how best to communicate risk information to patients 11 without causing undue anxiety or worry to patients.
Exploring ways to best communicate cancer risk to patients is important as the NHS cancer plan advocates effective communication between health professionals and patients, which can facilitate the delivery of high-quality care and empower people to be involved in decisions about their care. 12 While there is existing literature about communicating and sharing decisions with people who have cancer, 13 most communication evidence relates to people with existing cancer, but is not directly related to the use of cancer risk assessment tools designed for individuals with symptoms reporting to primary care who may not yet be aware of their cancer status. In fact, recent research in this area indicates cancer decision support tools which include QCancer are not being widely used. 14 The aim of this study was to explore the views of service users and primary care practitioners on how best to communicate cancer risk information when using QCancer, a cancer risk assessment tool, with symptomatic individuals in primary care consultations to enable them be involved in decisions on referral and cancer investigations.
QCancer was used as a reference tool because it has been validated externally. 8,9 2 | ME THODS

| Design and setting
We used a qualitative design employing semi-structured individual interviews and focus groups. The study took place in Lincolnshire in the East Midlands region of England during 2016. Ethical approval for the study was granted by the School of Health and Social Care Ethics Committee, University of Lincoln.
The eight stages of the Risk Analysis Framework 15 informed the development of the topic guide for the interviews, the analysis and interpretation of the data as well as the discussion of the findings.
The framework recognizes that a combination of some or all of the stages is important for effectively communicating risk to patients. 15 The eight stages are (to some extent ironically) stated as follows: (a) Get the numbers right; (b) Tell them the numbers; (c) Explain what the numbers mean; (d) Show them that they have accepted similar risks in the past; (e) Show them that it is a good deal for them; (f) Treat them nicely; (g) Make them partners; and (h) Combine all the stages. The Risk Analysis Framework was selected to inform the development of the topic guides (for the interviews and focus groups), the analysis and interpretation of the data, because the various stages of the framework served as a suitable theoretical background for explaining how best cancer risk can be communicated to a patient during consultations.

| Recruitment of participants and data collection
We recruited a convenience sample of service users and primary care practitioners (GPs and practice nurses) for the study. Prior to starting the individual interviews and focus groups, participants were asked to give written consent and were assured that they were free to discontinue the interview or focus group at any point. With permission of the participants, the interviews and focus groups were audio recorded and later transcribed verbatim. Notes were also taken which complemented the audio-recorded data.

| Data analysis
Data were transcribed verbatim and analysed using the framework approach 16  Further inductive codes were identified from the data. This involved the following steps. Two investigators (JNA and ANS) read and re-read transcripts to derive an initial coding framework which was discussed and agreed by the wider research team. Initial descriptive themes were developed by organizing codes. These were developed iteratively through further interpretation and discussion into a smaller number of overarching themes. For both service users and practitioners, data were collected and analysed until saturation was achieved; that is, no new codes or meaning were identified. 17 Data from service users and practitioners were analysed separately and then compared to determine the extent to which these views agreed or differed.

| Patient and public involvement
The research questions, the study design and interview guides were informed through discussion with the Healthier Aging Patient and Public Involvement (HaPPI) group at the University of Lincoln. Members of the HaPPI group also helped with recruitment of service users, helping to identify and distribute the study flyers to interested participants.

| Participants
We interviewed 36 participants, 19 service users (aged 21 to 71 years) and 17 practitioners (aged 33 to 55 years) as presented in Table 1. Of the 19 service users, two had a previous diagnosis of cancer, and the rest had relatives or friends who had a previous diagnosis of cancer.

| Themes
The analysis produced four overarching themes which addressed the research question on how best to communicate cancer risk information to patients when using a cancer risk assessment tool.
Participants felt that effective communication of cancer risk information between patients and practitioners when using cancer risk assessment tools was important, and that this could be enhanced in the following ways: personalizing risk information; involving patients when using the tools; being open and honest; and providing time for listening, informing, explaining and reassuring patients in a professional manner. Practitioners generally agreed with this view that personalizing risk could enhance patients' understanding of risk information, but some practitioners felt that the icon arrays used in QCancer could be improved, by arranging them together in rows rather than scattered throughout the diagram (see Figure 1). In line with this, one practitioner said: feel informed or understand the information, which could make them more anxious, less reassured, and they may lose trust in the clinician.

| Theme 1: Personalizing risk information
Practitioners explained their approach to informing patients.
One stated:

| Theme 3: Being open and honest
Being open and honest relates to the risk communication framework

| Main findings
We found a range of ways of improving how cancer risk information could be communicated to patients during primary care consultations. These included the following: personalizing risk information arising from the use of a cancer risk assessment tool; informing and involving patients when using the tool; sharing risk information honestly; and providing sufficient time for listening, explaining and reassuring patients in the context of a professional approach. Some findings contradicted others; for example, uncertainty about risk led to reluctance to openly and honestly share information.

| Comparison with existing literature
This study adds to previous evidence on the importance of personalizing information to suit the educational level, cultural background and the general level of understanding of individual patients, 18 as many patients, including those with cancer, prefer information specific to them (such as that derived from their own medical records) rather than more general information. 19 It also In another systematic review investigating design features of graphs in health risk communication, 23 there was evidence of patients being more able to recognize proportions with part-towhole sequential icon arrays (ie icons arranged in an ordered pattern) 24 than randomly arranged icon arrays, 25 or jittered icons (ie small unsteady or difficult to visualize icons). 26 This may explain the dislike of random-arrangement arrays found in this study and also reported in a previous qualitative study. 27 Participants thought that scattering blue 'sad' faces gave the impression that they were more numerous, that the risk perceived was higher, and that better structured graphics would ensure a visual representation which was clearer and easier for patients to understand. The importance of such design features supports the notion that the design characteristics of an intervention can help its implementation. 28 The literature on communicating results of (asymptomatic) screening is also sparse but highlights the importance of timely and effective communication, preferably verbal and face-to-face (rather than by letter or face-to-face) on patient understanding although the findings for anxiety or worry were mixed. 29 Providing accurate information and sharing decision making overall increased uptake of screening behaviour. 30 The other key difference with screening asymptomatic individuals is the presumption that the Wilson-Jungner criteria 31 have been met for the screening test, whereas there has been limited evidence presented of the outcomes of cancer risk assessment tools in terms of their acceptability, feasibility, false negatives and false positives and survival or patient experience outcomes.
Although service users in our study did not express difficulty in understanding the QCancer risk information, difficulty for some patients in understanding numbers could also complicate risk communication. 32 Around 40% of high school graduates in one study could not perform basic numerical operations, such as converting 1% of 1000 to 10 out of 1000, presenting a major barrier to understanding or interpreting health statistics. 33 Physicians may also find statistical information difficult to interpret and explain, 34  includes delivering bad news and sharing required information in a professional and responsible manner. 35 We found no demographic differences in the views of either service user or clinician participants on telling patients the truth about their cancer risk, although a recent study found both similarities and differences in preferences of men and women for truth-telling and decision making. 36 According to Chen et al, men and women had similar views of wanting to know about their medical condition, direct and frank truthfulness, and assistance in decision making for treatment. Truth-telling differed by gender in the following ways: women wanted family members present for confirmation of diagnosis, whereas men did not; men preferred truth-telling for only key points of their cancer, whereas women wanted detailed information; and men did not want to know their survival period, whereas women wanted this information. 36 When using a cancer risk assessment tool, although a few service users did not mind if they were not informed and some practitioners

| Strengths and limitations
The data collection strategy was flexible allowing the use of individual interviews (which provided information from service users and practitioners) and focus groups which facilitated the collection of diverse and detailed information from practitioner colleagues who decided to meet in a group, rather than being interviewed individually.
Individual interviews provided time for clinicians to provide in-depth information, prompted by the interviewer and topic guide, whereas the clinician focus groups benefitted from interaction and discussion between participants.
A key strength of the study was that service user and practitioner perspectives could be compared to highlight areas of agreement and disagreement between them. Another strength of the study was the achievement of data saturation in codes (ie no new ideas expressed by participants) and meaning (ie all ideas expressed were understood during the interpretation). 17 We followed the Consolidated Criteria for Reporting Qualitative Studies 45 (see Table S1), to ensure transparency and trustworthiness in reporting our research.
One limitation was that all nineteen service user participants were of White British ethnicity. The lack of representation from ethnic minority people is a limitation because GPs could use cancer risk tools with these patients and their views would be relevant and important. Although study publicity was circulated widely, it is possible that people from ethnic minority groups did not see the advertisement, were unable to understand it due to language difficulties, or had less interest in participating in the study. Previous research suggests that members of minority ethnic groups are less likely to participate in research studies if they lack confidence in language or understanding of the topic being researched. [46][47][48] They are more likely to participate if they are approached with sensitivity and they perceive the study to be beneficial. 46,49

| Implications for practice and research
News about cancer risk could cause undue anxiety or worry to patients. The findings from this study will help primary care practitioners to appropriately communicate cancer risk information to patients and minimize patient anxiety and worries during consultations. This paper addresses the issue of how best to communicate cancer risk information to patients presenting to primary care with symptoms which could be those of cancer. Further research on other barriers and facilitators to implementation of cancer risk assessment tools, with particular reference to QCancer, should be conducted.

| CON CLUS ION
Communication strategies suggested by participants in this study could be used to enhance the discourse between patients and practitioners when using cancer risk assessment tools during primary care consultations. As recognized in the Risk Analysis Framework that informed this study, the different communication strategies may be combined as appropriate to enhance communication in this context.

ACK N OWLED G EM ENTS
We thank the service users and primary care practitioners who participated, the PPI group who helped with the recruitment of participants and members of the Community, and Health Research Unit who commented on the paper.

CO N FLI C T O F I NTE R E S T
There are no conflicts of interest for Akanuwe, Black, Owen or Siriwardena.

AUTH O R CO NTR I B UTI O N
Siriwardena had the original idea for the study. The study was designed by Akanuwe and Siriwardena, supported by Owen and Black.
Fieldwork and analysis was conducted by Akanuwe supported by Owen, Black and Siriwardena. Akanuwe wrote the first draft of the paper, and all authors edited and approved the final paper.
Siriwardena is guarantor for the paper.

E TH I C A L A PPROVA L
The study was approved by the University of Lincoln School of Health and Social Care Research Ethics Committee.

DATA AVA I L A B I L I T Y S TAT E M E N T
The anonymized qualitative datasets will be available on request through the University of Lincoln repository.