Delivering the unexpected—Information needs for PSA screening from Men's perspective: A qualitative study

Abstract Background Making decisions about PSA screening tests is challenging, as it requires both knowledge of the possible benefits and harms of screening and an individual assessment of the patient's values. Our research explores how much and what information men perceive to be necessary with regard to screening for prostate cancer. Objective To explore men's information and associated needs for decision making in PSA testing. Design Qualitative interview study. Setting and participants We interviewed 32 men (aged 55‐69) about their decision making on PSA screening following counselling with a Decision Aid at their GP’s or urologist's practice in Germany. Main Outcome Measures Men's expressed needs for decision making in PSA testing. Methods All interviews were transcribed verbatim and analysed by framework analysis. Results Comprehensive pre‐screening counselling is needed. For the men in our study, information about test (in)accuracy, the benefit‐harm balance and consequences of the test were relevant and surprising. Additional needs were for interpretation support, a take‐home summary and time for deliberation. For several men, their physician's attitude was of interest. After being well‐informed, most men felt empowered to make a preference‐based decision on their own. Discussion Men were surprised by what they learned, especially regarding the accuracy and possible harms of screening. There is large variation in the breadth and depth of information needed, and some controversy regarding the consequences of testing. Conclusion and patient contribution A core set of information should be offered before men make their first PSA screening decision. Information about biopsy and associated side‐effects could follow in a short form, with details only on request. Knowledge about a high rate of false‐positive test results beforehand might help men handle a suspicious test result.


| INTRODUC TI ON
Until some years ago, and in part to date, cancer screening has been seen almost exclusively positively and uncritically, both in the community 1,2 and also by many physicians. Early diagnosis of cancer by screening seems to be generally perceived as advantageous, as this allows for early treatment. While the benefits of screening outweigh its possible harms for some types of cancer, this is debatable with regard to screening for prostate cancer (PCa) by testing for prostate-specific antigen (PSA). PSA screening seems to slightly reduce PCa-specific mortality, by 1 death per 10.000 person-years, [3][4][5] but is associated with psychological or physical consequences. The ethical dilemma of screening is that asymptomatic people are at risk of receiving a diagnosis of cancer which would never have affected them in their life-so-called overdiagnosis, often associated with overtreatment and corresponding side-effects. 6,7 In Germany, PCa screening is common in general practice and ambulant Urology clinics. 8 However, costs are not covered by the statutory health insurance. Men often undergo the test with a lack of knowledge, in particular with little understanding of the unclear benefit and, furthermore, the potential repercussions of the initially innocuously perceived blood test. 9 The men tend to an unbalanced perception and overestimated screening benefits. 10,11 Due to the controversial benefit-harm balance, many countries around the world have decided against opportunistic mass PSA screening, advocating for an informed, values-based, individual decision. [12][13][14] The combination of adequate knowledge and of patients' values comprises the foundation for a 'good decision'. 15,16 Such a good decision is difficult to achieve and patients should neither be left to make a decision nor 'forced' towards one. Instead, patients should be involved in Shared Decision-Making (SDM). 6,14 However, we still observe a lack of SDM in clinical encounters. [17][18][19] Decision aids (DA) may be used within a consultation to enhance SDM. [20][21][22] In situations where more than one reasonable option is available, DAs make explicit that a choice exists. They present benefits and harms, optimally supported by pictographs or icons which communicate risk graphically. 20 Nevertheless, physicians and experts disagree on the amount of information that should be given prior to the screening test.  21,22 Furthermore, communication goals differ, ranging from convincing men to be screened or not to be screened to facilitating and supporting patient choice. 23 The approach to defining DA content from an objective, scientific point of view is also open to consideration, but does not necessarily coincide with men's needs.
This raises the question of which support should be provided to men. 24 There are some quantitative studies about PSA testing to determine physicians' rating of the importance of key facts, 21 to explore which information has an effect on men's interest in screening, [25][26][27] or to examine their knowledge. 2,11,28 Also, there are qualitative studies asking patients, 29,30 GPs, 31,32 a community jury, 33 or experts and patients 22 about PSA screening. These studies were conducted in Australia, the UK and the United States. Therefore, we aimed to assess the transferability of previous findings to the German setting. Moreover, in exploring the needs of men regarding decision support on PSA screening, needs can imply, in addition to factual information, other decision support. 15,24 A further focus lies in the degree to which potential (treatment) consequences should be presented obligatorily-for PSA 34 and for other screening decisions 35 this is still in debate. Owing to the abovementioned problem, we chose a special setting. We interviewed men who had been guided through a structured DA. We aimed to extract practical recommendations to help providers in counselling men adequately and in supporting shared decision making.

| Design
The multiphased PSA-inform project consisted of the development and user evaluation of a transactional DA for PSA screening (arriba-PSA), 36 followed by a later RCT 37 to, among other things, determine the influence of the transactional DA on men's decisional conflict.
The results presented here are part of the early qualitative user evaluation after the ODSF-oriented 15 development of the transactional DA for PSA screening. 38 Within this context, we took the opportunity to ask men directly about their own needs regarding clinical counselling. Men were guided by their physician through a sequence of information and deliberation steps following a specified course of talk and graphs irrespective of what they needed to know. Of course, men had the opportunity to ask questions and to obtain additional explanations. The content of the DA, however, was fixed in advance. This provided a unique opportunity to study men's information needs. We assume prior counselling with a DA as prerequisite to evaluating informational needs, as one can hardly assess the salience of knowledge one does not have.

K E Y W O R D S
consumer health information, counselling, decision aid, prostate-specific antigen (PSA), qualitative research, shared decision making

| Setting and recruitment
GPs in our research network and urologists in a district of Hesse in Germany were invited to participate in the pilot evaluation of the DA.
For counselling, the GPs and urologists attended a two-hour training session which focused on PSA screening and the application of the consultation with the DA. Afterwards, each physician recruited three to four men, aged 55-69, representing the core age group of ERSPC, 39 from their daily consultations. For inclusion, men should have had a preventive check-up examination or raised the question of PCa screening, respectively. All men had to be asymptomatic regarding the prostate and had to have sufficient German language skills. We included men with and without prior PSA test experience.
A sample of 8-10 physicians and of about 30 men was deemed appropriate to reach diversity in characteristics and to reach thematic saturation. 40,41 Sample and recruitment considerations were based on the DA development and evaluation study (publication in preparation). Our study was approved by the Ethics Committee of the Faculty of Medicine, University of Marburg (72/13). All participants, men and physicians, gave written informed consent.

| Intervention -the DA
The decision aid (arriba-PSA 36 ; see Appendix S1) contained informa- Benefits (here defined as reduction of PCa-related morbidity and mortality 6 ) and harms (here defined as false-positive results, psychological harm, complications of prostate biopsy, overdiagnosis, overtreatment, or side-effects of treatment 6 ) were not labelled as 'benefit' or 'harm' in the DA, as the interpretation has to be made individually. Accordingly, the DA ends with a summary and a handout with a short value clarification exercise.

| Data collection and analysis
Individual, semi-structured face-to-face interviews with men as key informants were conducted by the study team (KK, KS, MB, CCA) between 1-2 weeks after the men had received counselling with the DA. This time span was chosen to allow men recollection of the DA content for the interview. If necessary, the DA slides were shown again but not discussed within the interview to check for comprehension. All interviewers are experienced health professionals and researchers and participated in a shared training of the interview guide (extract see Appendix S2). All interviews were audiotaped, pseudonymized and transcribed verbatim.
We used the Ottawa Decision Support Framework (ODSF) to inform the guideline and to map our findings. Following the ODSF, we clustered the needs of men in three domains (gaining knowledge, getting support and becoming aware of own preferences). 15,16 All interviews, originally focused on evaluating the DA, were fully coded.
Afterwards, all interviews were recoded and re-analysed with regard to the focus of information needs. Only those statements which were given directly as an answer to a question that focused on decisional needs or which were markedly accentuated by the interviewee, for example by adjectives or frequent iterations, were eligible for this secondary analysis. We derived the first coding scheme deductively, based on the interview guide and the ODSF, and augmented it inductively according to emerging themes.
Our analysis was based on the framework approach. 42,43 Specifically, we sorted recurring or important themes into a cate-

| Participants
Nine physicians (7 GPs, 2 urologists) and 32 men participated in the study. Due to the technical loss of one interview, we analysed the interviews of 31 men. The average interview length was 29 minutes (range 8-56 minutes; SD 12.2; IQR 17.6). Half of the men reported having had a prior PSA test. The characteristics of study participants are summarized in Table 1 and Table 2, respectively.

| Decision-making needs and decision support
Although nearly half of the men had already had one or more PSA tests before, they were surprised by what they heard and learned in the specific setting of this study.

| Clarify decision and needs-Need for prescreening counselling
One of the main results is that the men stated that they aim to be thoroughly informed. For them, it should be made clear that they have options (to test-not test/postpone decision) and that an individual decision should be based on their own values and preferences. Several men noted that they either had not been informed before having a PSA test or that they lacked comprehensive and neutral information. In contrast, two men stated that for them the test result is the only relevant information. They would not need any information on test characteristics. After participating in this study, two men felt that their doctors should have been more knowledgeable regarding PSA testing and its consequences.

| Knowledge and expectationsinformation needs
Information needs were discussed according the structure of the DA: information about the prostate and PCa, test (in)accuracy, benefitharm balance, possible consequences of the PSA test and of overdiagnosis/overtreatment. With regard to facts men ought to know, we identified several consistent key themes. However, men differed regarding the information they found most important for their decision making. Nevertheless, we identified three issues important for most men: the test (in)accuracy, the benefit-harm balance and Test accuracy ✓ ✓ ✓

Consequences
In detail (e.g., biopsy) ✓ ✓ Contr. awareness of subsequent procedures/consequences. The amount of overdiagnosis was seen as essential by some patients, while test consequences were seen controversially. All of this information also took men by surprise. Table 3 summarizes men's needs (✓) and controversial needs (contr.) and compares them with the recommendation and the DA used.

Awareness of Consequences
During the interviews, we learned that the men had little prior knowledge of the prostate, PCa and PCa screening. Nearly all of the interviewees perceived information on incidence and mortality as helpful in understanding that PCa is a common condition in older men, but the mortality is quite low. However, there was a tendency to mistake the incidence and mortality of PCa. man) The concepts of overdiagnosis and overtreatment were completely new for most men. Living onwards with a known diagnosis was felt to be very burdensome. In a few men this information was the crucial point for their decision making.

| Support and resources-additional needs
In addition to facts about the PSA test itself, men highlighted further needs regarding support for decision making (see Figure 1).

| Discussion
Within this study, we used the unique opportunity to explore PSA screening information and further needs of men who were counselled with a structured DA. First of all, the results revealed a large variation regarding the breadth and depth of information men need to know. Moreover, in the specific setting of this study, men were highly surprised by what they learned. While our findings mostly correlate with those of previous studies, 21,29,33,45,46 we examined the transferability for the German setting.
Our study provides further evidence for a general need in men for comprehensive, balanced information and a clarification that an individual, values-and preference-based decision is to be made. In particular, this needs to take place before the decision is made the While men agree on the necessity of nearly all the themes mentioned above, albeit with individual significance, there are differing views regarding two themes (test accuracy and overdiagnosis).
This includes the amount of detail on the consequences of a suspicious PSA test. Nevertheless, giving short information on potential consequences and side-effects before testing could influence the screening decision. 46 Several men expressed a demand for such information, although there was a tendency to make the decision step by step. Other studies support our findings. 33,50,51 Within our results, two further aspects are worth mentioning.
First, designating a test result as 'positive' or 'negative' could be misinterpreted by men; a less ambiguous wording would be 'suspicious' and 'unsuspicious', respectively. A second source for misunderstanding was identified regarding the interpretation of incidence and mortality.
In addition to factual information, further support-related needs were mentioned. Physicians were seen as the preferred source of information and were expected to interpret available information.
Particularly for previously uninformed men, the amount of new, surprising and conflicting information could lead to increased uncertainty and therefore could arouse decisional conflict. To make up their mind, men asked for information materials they could take home in order to digest what they had learned; for the same reason, they felt they needed some time to consider the information and not feel pressured to make an immediate decision. These results correspond with those of other studies. 29,33,50,52,53 Another area of disagreement was found regarding the wish to hear the physician's attitude. Many, but not all men, were interested in this. The reason could be age-related, that is in the generation of men aged 55 years and above, many wanted to know their phy-

| Strengths and weaknesses
This study provides a valuable insight into screening information needs as a prerequisite to informed decision making for asymptomatic men. Our analysis regarding information needs was deductive, and further issues were also derived inductively.

| Conclusion
Men were highly interested and grateful for the education they received by structured counselling on the basis of a DA. Regarding information, they agree on the absolute necessity of most themes.
Nevertheless, we found large variation regarding the breadth and depth of information needs. Many of the issues broached caused surprise; at least, cursory information on test accuracy and unknown constructs such as overdiagnosis were considered necessary. In addition to information, other needs became apparent, such as interpretation support, an information leaflet to enable reflecting on the information and time for making the decision. In any case, merely a recommendation or having the test without any or only selective, biased information is not what men expected.

| Practice implications
Even if men had undergone screening tests several times, their baseline knowledge differed, and many were not aware that these tests can cause at least some harm. After consultation with a Decision Aid, unbiased, newly impressive and in part surprising information arose. This calls for a core set of information which should be presented to each man. Optimally, this should be done before a man's first screening decision, as we know that the strongest predictor for having a PSA test is having previously done PSA tests. 46 The presentation should include information about options, benefits and harms, including overdiagnosis, uncertainty and that further steps and associated side-effects could follow in the event of a suspicious test result in short form. The degree of detail, but not the core components, could be adapted with regard to interest, knowledge and intellect, but should not be anticipated subjectively beforehand. This semi-individualized approach is not completely standardizable, but rather underpins the need for personal counselling, paying attention to a man's reaction to the core information provided. Advance knowledge about the possibility of false-positive test results puts a suspicious test result into perspective.

ACK N OWLED G EM ENTS
We would like to thank all of the physicians and men who participated in this study. Open access funding enabled and organized by projektDEAL.

CO N FLI C T O F I NTE R E S T
KK, KS, CCA and MB have no conflicts of interest to declare. NDB is Co-CEO of the 'Gesellschaft für Patientenzentrierte Kommunikation' (Organization for Patient-centred Communication). This is a registered non-profit entity contracting with health insurers and providers aiming at the dissemination of arriba decision support software.
He obtains no financial revenue from this organization apart from travel expenses.

AUTH O R CO NTR I B UTI O N S
NDB designed the study, KK, KS, CCA and MB participated in the acquisition and analysis of data. All authors discussed and interpreted the data and the results. KK drafted the manuscript; KS, CCA, MB and NDB revised it critically; and all authors have approved the final article.

I N FO R M E D CO N S E NT A N D PATI E NT D E TA I L S
We confirm that all patient/personal identifiers have been removed or disguised so the patient/person(s) described are not identifiable and cannot be identified through the details of the story.