Appraising risk in active surveillance of localized prostate cancer

Abstract Objectives Men diagnosed with low‐risk prostate cancer are typically eligible for active surveillance of their cancer, involving monitoring for cancer progression and making judgements about the risks of prostate cancer against those of active intervention. Our study examined how risk for prostate cancer is perceived and experienced by patients undergoing active surveillance with their clinicians, how risk is communicated in clinical consultations, and the implications for treatment and care. Method Participants were nine patients and three clinicians from a university hospital urology clinic. A staged, qualitative, multi‐method data collection approach was undertaken, comprising: observations of consultations; patient and clinician interviews; and patient surveys. The three data sets were analysed separately using thematic analysis and then integrated to give a comprehensive view of patient and clinician views. Results Thirty data points (eight patient surveys; 10 observations of consultations between patients and clinicians; 10 patient interviews; and two clinician interviews) combined to create a detailed picture of how patients perceived and appraised risk, in three themes of “Making sense of risk”, “Talking about risk” and “Responding to risk”. Conclusion Effective risk communication needs to be finely tuned and timed to individual patient's priorities and information requirements. A structured information exchange process that identifies patients' priorities, and details key moments in risk assessment, so that complexities of risk are discussed in ways that are meaningful to patients, may benefit patient care. These findings could inform the development of patient‐centric risk assessment procedures and service delivery models in prostate cancer care more broadly.

The decision to take part in an AS protocol involves patients, in discussion with clinicians, weighing the risks of regularly monitoring of PCa against those of active treatment, based on the clinician's judgement. The risks of active interventions can include serious side-effects impacting a patient's bladder and bowel continence, erectile function, fertility and hormone balance, 11 which then require management from multiple health services.
Nevertheless, AS protocols are not without risk of negative effects. Patients may be asked to regularly undergo invasive and discomforting biopsies, ultrasounds, and expensive imaging procedures, such as multi-parametric MRI scans. 12,13 Moreover, as it does not involve the eradication of the cancer, AS may be associated with anxiety and uncertainty. 14,15 Patients' perceptions of risk and their certainty of the effectiveness of treatments for PCa are influenced by their understanding of clinical outcomes, 15,16 in turn affected by a number of psychosocial factors, such as family support, clinician communication and attitude, 16 and anxiety and distress. 7 In this study, we sought to uncover the experiences of men with low-risk PCa responding to the risks associated with their cancer.. We aimed to identify risk from the perspective of men with localized PCa, who were considering or using an AS protocol to manage their risk of cancer metastasizing, and from the perspectives of their treating clinicians. We asked: (a) "How do patients perceive and experience risk for PCa; and how is risk communicated between clinicians and patients'?" and (b) "What are the implications of patients' perceptions of risk on their treatment and care?" In this study, we define risk as the probability of a particular clinical event occurring, 17 such as the risk of PCa metastasizing. Risk is typically expressed to patients in terms of numbers, 18 to describe probabilities, including percentages of patients who may experience side-effects from curative treatments, or the likelihood of death at 10 years for those on particular treatments. Even though risk is expressed in terms that are clinical (impacting on a person's physical health) patients can also experience risk in ways that are non-clinical (impacting on their quality of life) or in terms of psychosocial implications (impacting on their relationship and mental health). 16 Risk communication in this study builds on our previous work in understanding how risk is discussed with patients, and expressed, in breast cancer. 17,19 2 | ME THOD

| Setting
The project was conducted in an Australian university hospital urology clinic, between November 2016 and December 2017. The clinic is part of a private practice outpatient service, attached to the university hospital, serving the needs of a diverse patient demographic and offering wide-ranging clinical services. Ten Urologists see patients in their clinic rooms, and while the senior clinician practises solely at the university hospital clinic, the remaining nine Urologists see patients at multiple clinic sites. General practitioners (GPs) refer patients directly to the Urologists as private practitioners, rather than to the urology clinic. As such, each Urologist's caseload is separate to that of the other Urologists. The Urologists attend a fortnightly multidisciplinary team meeting to discuss their patients collaboratively and to examine patients' cases in detail. The patient flow pattern of the clinic is shown in Figure 1.
As the clinic is a private practice, in this study, the men who used the clinic services were generally from a middle socio-economic F I G U R E 1 Urology clinic patient pathway for assessment and treatment GP referral to Urologist PSA testing MRI Biopsy Urology clinicdiscuss results

Treatment decision and plan
band, living in metropolitan areas and surrounding districts, with a small number travelling up to 2 hours to attend the clinic. The men were typically aged from mid-40s to 70s and came from predominantly white or south-east Asian backgrounds. There was a mix of working and retired men, from occupations such as business and trade.

| Recruitment
The study was advertised using flyers placed in the Urology clinic waiting room area, to ensure that patients visiting the clinic were aware that the research was taking place. In addition, patients considered suitable for the study (that is, those who were clinically appropriate candidates, who had PCa, were under or considering AS, and showed no psychological sign that involvement would raise anxiety levels or lead to any undue harm) were identified by the two par-

| Data collection
A multi-method, staged approach was used to gather data, 21 comprising: patient surveys; observations of consultations between patients and clinicians; and individual interviews with patients, and with clinicians. Data collection was planned to be consecutive with one stage following the next, so that surveys would precede observations, and be followed by interviews. However, this was not always feasible, as some patients were referred to the study after their consultation appointment, while others were not available to participate in an interview on the same morning as their clinic appointment. Thus, the researchers conducted each stage of the data collection procedure consecutively yet at the convenience of the participants.
The patient survey was purpose designed for the study, based on information gained through discussion with the Urology Clinic team.
Survey questions covered demographic and health-care questions (Appendix S1: Survey questions). Data collection was opportunistic, with surveys completed prior to observations or more commonly at the start of interviews. Moreover, discussion of the survey ques-

| Data analysis
The analysis process was systematic and iterative, with codes derived from the data and involving contributions from multiple data sources and several members of the team. The three forms of data, in line with a staged approach to data capture, 23 were first analysed separately. Once complete, the data sets were integrated as one combined data set, to provide a more comprehensive view of patient and clinician perceptions and experiences.

| Survey data
Participant data from the demographic surveys were compiled, analysed descriptively and summarized (Table 1).

| Observation data
Observations of consultations between patients and the two Urologists were undertaken. Eight of the consultations were audiorecorded and transcribed. Two patients requested that only field notes of their consultations were taken and that the consultations were not audio-recorded. Observation data were read through by the researcher who had attended the appointment, to identify the relationship between observation notes and interview data and how observations added depth of understanding to interview data, in particular, around risk communication between clinician and patient.
Observation findings were assessed thematically, and key themes were merged with themes derived from interviews to form a richer narrative of the consultation that was representative of the interactions between the Urologists and patients.

| Interview data
Data were analysed in a stepwise procedure to ensure rigour. The researchers then met to discuss the codes that emerged and agreed on an initial thematic coding framework 24 that could be applied to code the remaining transcripts. Patient and clinician interview transcripts were then uploaded into NVivo 11 (QSR International) to manage the data efficiently and to assist with the understanding of theme development. Once all the transcripts were read as a complete data set, additional codes that were not captured in the initial sample were identified. The coding framework was discussed by the two researchers and amended to complete the preliminary analysis. One researcher grouped the codes to form subthemes as they emerged from the data. For example, the code "clinician's risk representation" formed part of the subtheme "communicating risk to patients" and then rolled up into the overarching theme of "talking about risk". The themes and subthemes were discussed within the wider research team (comprising health-care professionals and academic researchers) using tried and tested group work activities 25 until consensus on key themes emerging from the thematic analysis was reached.

| Combined data
Members of the study team reviewed the analysis of each data set and discussed the meaning of the combined findings. Ongoing discussions ensued until a final thematic framework emerged across all study stages. This was supported by the corroboration of observation findings with interview findings, alongside the patients' priorities for their health and well-being, indicated in the surveys. These combined data sets were reanalysed for agreement or disagreement between data sources. Minor inconsistencies across data sets were discussed and addressed early in the process, leading to a more comprehensive set of findings, for deeper understanding of the data and consideration of further research questions. 26,27

| RE SULTS
In total, 30 sources of data were collected and compiled: eight patient surveys; 10 observations of consultations between patients and clinicians; 10 interviews with patients; and two interviews with clinicians (one Urologist and the specialist Prostate Cancer Nurse, Table 2).
Nine men (of 10 approached) consented to take part in the study.
Eight men with a diagnosis of localized PCa, who were using or had considered an AS protocol, were enrolled. Of these, one man had previously undergone surgery to remove a high-risk cancer and was now undertaking AS to monitor any recurrence. The ninth participant did not have a PCa diagnosis, but had a strong family history of prostate cancer, who regularly undertook screening tests for onset of the disease that was akin to the AS protocol used in the clinic (see demographic data below for more detail on the patient cohort).
Clinician participants were two male Urologists (four Urologists were approached) and a male specialist Prostate Cancer Nurse. The three clinicians participated in different data collection activities.
The two Urologists participated in observations of consultations, but no consultations with the specialist Prostate Cancer Nurse were observed. One Urologist and the specialist Prostate Cancer Nurse took part in interviews.
Of the nine patients considered suitable for an AS protocol, seven remained on AS, one opted for surgery and another for radiotherapy. Of the consultations, six were with one Urologist, and four were with the second Urologist. Patient consultations with the specialist Prostate Cancer Nurse were not observed.
Combined data generated a detailed picture of the way patients and clinicians perceived, discussed and responded to risk in this care setting. Disagreement between data findings was minor and therefore is not presented; for example, the number of patients reporting using friends as an information source differed slightly between survey and interview data. Three themes integrated the findings to represent patients' appraisal of risk. These were "Making sense of risk", "Talking about risk" and "Responding to risk". Quotes illustrating each theme are placed in tables and referenced to the related text. Participants are identified as either patient (P) or clinician (C).   One patient with a family history of PCa (father and grandfather) drew on their collective experiences to motivate vigilance of his prostate health and to keep himself up to date with detection and treatment procedures (Table 3, quote 9). Patients also drew on information given to them by their clinicians during consultations. This is outlined in the following section, describing how risk was discussed in clinical consultation.

| Talking about risk
During consultations, clinicians used a patient-centred approach to identify the most appropriate prostate cancer intervention for each patient. This approach was evident in the use of shared decision-making principles (such as discussion of evidence-informed strengths and weaknesses of the treatment options, with reference to the patient's values and preferences). The approach was put into practice to support patients to reach their decision from the way information was presented to the patient, and the interaction between patient and clinician, both verbal and non-verbal.
Both Urologists adopted a similar approach to patient care, and the body language observed between patients and Urologists indicated that patients seemed relaxed and comfortable through-

| Responding to risk
As patients gained understanding of the risks they faced, they sought ways to take control over their health. This frequently took the form of self-management approaches to minimize risks, reduce anxiety and uncertainty, and determine their way forward (

| D ISCUSS I ON
The findings from this study add to previous research in PCa care that has investigated patient decision making in PCa diagnosis and treatment, 16 such as men's understanding of benefits and harms TA B L E 5 Responding to risk: Participant quotes associated with PSA testing 28 ; and their decision-making and treatment preferences. 29,30 However, in this study, we have concentrated on risk, through qualitative interviews, observations and surveys.
The evidence in our study suggested that effective risk communication in PCa may benefit from fine tuning to individual patient's priorities and information needs. Based on evidence from observation of consultations and patient and clinician interviews, we propose a structured approach to information provision that emphasizes that informed decision making about interventions for PCa is a multilayered exchange of information. 31 In this way, our findings build on previous recommendations for communicating risk in cancer, identifying helpful modes 32,33 and structures 31,34 for risk communication.
Clinicians in this study demonstrated that they balanced clinical judgement with evidence-based information, to deliver a patient-

Information for clinicians
During their first clinic appointment, patients could be asked to complete a proforma (written or verbally) to gather information about their priorities, information needs and expectations of the clinic service. This information could be provided to the Urologist and specialist Prostate Cancer Nurse, to have patients concerns documented and indicate areas for discussion during the initial clinic appointment.

A. About the clinic service
Prior to the first clinic appointment, patients should be sent information about the urology clinic service. This could include details about the services provided by the clinic, including the specialist Prostate Cancer Nurse service; hours of operation and availability of the clinic staff; and contact numbers for the clinic. Additionally, some patients will have psychosocial needs that may be beyond the scope of the clinic to address. To ensure that patients' psychosocial needs and related risks are fully assessed, details of a counselling service linked to the clinic should be included in the information given to patients.

B. About how risk is calculated and expressed in PCa clinical settings
As some patients are clearly unaware of, or do not fully understand, clinical risk classifications (low-, medium-and high-risk), and how the risk classification applies to them, patients would benefit from receiving written information about how risk is formally assessed and how Gleason scores are derived, including diagrams that explain what low-, medium-and high-risk classification means.

C. About PCa and treatment options
Following the initial appointment after biopsy, or whenever it is indicated that a patient is suitable to be considered for an AS protocol, patients should be provided with information about active surveillance in PCa, and the benefits and disadvantages of AS and active treatments. Patient decision support tools exist for these reasons; however, there is a need for up-to-date tools that are user-designed 38 and specific to the Australian clinic context. 35,39

D. About self-management and risk minimization
Information should be created for all urology clinics to be handed to patients upon attendance that can support patient self-management, that is, evidence-informed health literature to address the information gaps identified by the participants in our study, including positive health behaviours and potentially harmful behaviours to avoid.

Review
For the purpose of review, patients' initial concerns that are discussed with clinicians could be formulated into a list for patients to reflect back on before they attend follow-up review appointments. This would allow patients to consider what had initially concerned them, follow-up on any issues further and remove any old issues that were satisfactorily addressed. It would also enable patients to raise new concerns, or discuss changes to their life circumstances, that may have arisen over the review period. Compliance with long-term monitoring of PCa, where patients fail to return for clinical review after initial consultation and agreement of a treatment path, is an ongoing concern to urology clinicians. 8,31 Embedded within clinics' patient review processes, engagement with an AS protocol may be sustained if patients consider their early interactions with the Urology clinic have addressed their needs and that their prostate health will be expertly, and proactively, monitored.
To optimize engagement with the clinic service, the information gaps 36 and psychosocial needs 7 indicated by patients in this study could be addressed by the introduction of a structured process to gather, utilize and disseminate information within routine clinical care, and incorporate patients' recommendations for more effective risk communication with their clinicians. As described above, clinicians in this study communicated with their patients within the clinic appointment in a way that supported shared decision making.
To build on this strength, we propose the use of an information exchange process to ensure patients and clinicians receive appropriate, well-timed information in an accessible format to assist their decision making. As the specialist Prostate Cancer Nurse already provides patients and families with health literature and links to PCa support services, this process may be best coordinated through the specialist Prostate Cancer Nurse service. To our knowledge, this study is the first to derive a structured process of information exchange for treatment of low-risk PCa from the combined perspectives of patients and clinicians. Table 6 offers brief recommendations based on our findings. The effectiveness of such an approach requires evaluation.

| Limitations
Although patient recruitment was limited to patients considered suitable for AS by the two participating Urologists, there was variation within the patient group. Not all patients had PCa, and not all were using AS. Nevertheless, this variation resulted in more complex data, deepening our understanding of how patients appraise risk and make decisions based on this appraisal. Time frame sampling was a strength in accessing a more naturalistic sample, but because of the imposed time constraint, limited the sample size.
This study has generated a hypothesis about effectiveness rather than proving effectiveness with a small participant cohort.
Evaluation of the effectiveness of the approach we are suggesting has the potential to optimize engagement between health-care professionals and patients and amongst health-care professionals, leading to greater shared care and shared decision making within the consultation. While this is currently missing, evaluation of effectiveness would form an invaluable element of further research, using both quantitative, demonstrable measures and qualitative, in-depth assessment to clarify the use and value of a new information exchange system for all concerned.
The study took place in a single private health clinic using a small patient and clinician cohort. Patients were from Anglo-Australian backgrounds and well-educated, which may reflect those patients interested in research participation. Only one Urologist participated in an interview, preventing comparison between interview the Urologists involved. Consultations with the Prostate Cancer Nurse also were unable to be observed. As such, the findings are unable to be generalized to larger or publicly funded urology services. Nevertheless, we have gained rich and extensive insight into how men from diverse age groups and life stages appraise risk and identify their priorities and information needs, which we think have scope for other urology studies.
Female researchers interviewed male patients and clinicians in a specifically male health area. It is unknown whether this had any impact on recruitment to the study or the type of information participants were willing to divulge.

| CON CLUS ION
Men living with localized PCa, using an AS protocol to manage their condition, appraise risk in interrelated ways. Understanding of the risks they face, and awareness of their personal priorities, formed the basis of discussions with their clinicians. Our findings that patient's appraisal of risk was influenced by their priorities and information sources indicate that effective risk communication in PCa could be enhanced by well-timed provision of information aligned to patients' priorities, to reduce anxiety and uncertainty. Patients and clinicians could benefit from a structured process to identify patients' priorities prior to consultation, where the complexities of risk are discussed in ways that are meaningful to them. Additionally, access to a specialist Prostate Cancer Nurse contributed to the quality of, and satisfaction with, the service offered by the clinic, providing a valuable model of service delivery incorporating patient-centric risk assessment procedures and care. By explicitly addressing patients' needs, patient engagement with the clinic service may be enhanced over time.

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interest to declare.

E TH I C A L A PPROVA L
Ethical approval for the study has been granted by Macquarie University Human Research Ethics Committee (HREC), approval number 5201600638. Knowledge translation will be achieved through publications, reports and conference presentations to patients and families, health professionals and researchers.

DATA S H A R I N G S TAT E M E N T
Data available on request from the authors.