Selecting men diagnosed with prostate cancer for active surveillance using a risk calculator: a prospective impact study
Heidi A. van Vugt, Erasmus MC, Room NH-227, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. e-mail: firstname.lastname@example.org
Study Type – Prognosis (cohort series)
Level of Evidence 2a
What's known on the subject? and What does the study add?
The present study is one of the first to investigate urologists' and patients' compliance with recommendations based on a risk calculator that calculates the probability of indolent prostate cancer. A threshold was set for a recommendation of active surveillance vs active treatment.
Active surveillance recommendations based on a prostate cancer risk calculator were followed by most patients, but 30% with active treatment recommendations chose active surveillance instead. This indicates that the threshold may be too high for urologists and patients.
- • To assess urologists' and patients' compliance with treatment recommendations based on a prostate cancer risk calculator (RC) and the reasons for non-compliance.
- • To assess the difference between patients who were compliant and non-compliant with recommendations based on this RC.
PATIENTS AND METHODS
- • Eight urologists from five Dutch hospitals included 240 patients with prostate cancer (PCa), aged 55–75 years, from December 2008 to February 2011.
- • The urologists used the European Randomized Study of Screening for Prostate Cancer RC which predicts the probability of potentially indolent PCa (P[indolent]), using serum prostate-specific antigen (PSA), prostate volume and pathological findings on biopsy.
- • Inclusion criteria were PSA <20 ng/mL, clinical stage T1 or T2a–c disease, <50% positive sextant biopsy cores, ≤20 mm cancer tissue, ≥40 mm benign tissue and Gleason ≤3 + 3. If the P(indolent) was >70%, active surveillance (AS) was recommended, and active treatment (AT) otherwise.
- • After the treatment decision, patients completed a questionnaire about their treatment choice, related (dis)advantages, and validated measurements of other factors, e.g. anxiety.
- • Most patients (45/55, 82%) were compliant with an AS recommendation. Another 54 chose AS despite an AT recommendation (54/185, 29%).
- • The most common reason for non-compliance with AT recommendations by urologists was the patient's preference for AS (n= 30). These patients most often reported the delay of physical side effects of AT as the main advantage (n= 19).
- • Those who complied with AT recommendations had higher mean PSA levels (8 vs 7 ng/mL, P= 0.02), higher mean amount of cancer tissue (7 vs 3 mm, P < 0.001), lower mean P(indolent) (36% vs 55%, P < 0.001), and higher mean generic anxiety scores (42 vs 38, P= 0.03) than those who did not comply.
- • AS recommendations were followed by most patients, while 29% with AT recommendations chose AS instead.
- • Although further research is needed to validate the RC threshold, the current version is already useful in treatment decision-making in men with localized PCa.
European Randomized Study of Screening for Prostate Cancer
potentially indolent PCa
12-item Short-Form health survey
State Trait Anxiety Inventory
Memorial Anxiety Scale for Prostate Cancer
Decisional Conflict Scale
Center of Epidemiologic Studies Depression scale, EPQ, Eysenck Personality Questionnaire
Prostate Cancer Research International: Active Surveillance
The incidence of potentially indolent prostate cancer (PCa) has risen the last two decades, mainly as a result of PSA screening [1,2]. Autopsy studies show a high prevalence of these small, localized, well-differentiated tumours in men dying from other causes . Many of these cancers will remain non-harmful during a man's lifetime [4,5]. To avoid overtreatment they could be closely monitored with the aim of switching to active treatment (AT) with curative intent if progression occurs . Prospective analyses of men undergoing such an active surveillance (AS) strategy show favourable 10-year PCa-specific survival rates approaching 98% [4,6]. Crucial for a successful AS strategy is the reliable identification of indolent PCa; however, a key problem is that it is difficult to differentiate between men with aggressive localized PCa and indolent PCa. Prediction models have been developed to support the identification of indolent PCa [7,8], but the use of these models in urological practice is not standard.
We implemented levels three and six of the six levels of the risk calculator (RC) based on data from the European Randomized Study of Screening for Prostate Cancer (ERSPC) in five Dutch hospitals (http://www.prostatecancer-riskcalculator.com). Level three calculates the probability of a positive biopsy using serum PSA, outcomes of DRE and TRUS, and TRUS-assessed prostate volume. Level six calculates the probability of a potentially indolent PCa (P[indolent]) using serum PSA level, prostate volume, mm cancer tissue, mm benign tissue, and the Gleason score at biopsy. The present study addresses level six. As a rule for treatment decision-making we decided that AS would be recommended if P(indolent) was >70%, and AT would be recommended otherwise. This 70% threshold was based on a study where an existing clinical RC was validated and adapted towards a screening setting, resulting in a 94% sensitivity (actively treating important PCa) and a 32% specificity (applying AS to potentially indolent PCa) [7,8]. The RC performed well in a mixed screening/clinical cohort with an area under the curve of 0.77 .
The aim of the present study was to assess: (i) urologists' and patients' compliance with treatment recommendations by the ERSPC RC level six; (ii) the reasons for non-compliance; and (iii) the difference between patients who were compliant and non-compliant with AS and AT recommendations based on a RC.
PATIENTS AND METHODS
Eight urologists from five Dutch hospitals studied patients, aged 55–75 years, from December 2008 to February 2011. Before the start of the implementation project urologists and nurses were informed about the aim, use and interpretation of the outcome of the RCs. The nurses' role was the promotion of the use of the RC and collecting data.
Patients needed to fulfill the following criteria; biopsy-confirmed PCa, PSA level <20 ng/mL, clinical stage ≤ T2c disease, <50% positive sextant biopsy cores, ≤20 mm cancer tissue, ≥40 mm benign tissue and Gleason score ≤3 + 3. These patients did not participate in a screening trial. All patients provided written informed consent. The study was approved by the Institutional Review Board of the Erasmus Medical Centre, Rotterdam.
Urologists calculated P(indolent) using the RC level six and used this outcome in their treatment advice to their patients (Fig. 1) . The RC was based on sextant biopsy outcomes. When more than six biopsy cores were taken, mm cancer tissue and mm benign tissue were calculated pro rata ).
After the treatment decision was made, both urologists and patients received a questionnaire. Urologists were asked to indicate their own and patient's compliance with the recommendation by the RC and, if applicable, reasons for non-compliance. PSA level and the other necessary data for the use of the RC, the P(indolent) and the patient's final treatment choice were also recorded.
Patients were asked to indicate advantages and disadvantages of their treatment choice using open-ended items, with space for three possible responses. These were grouped and counted independently by the author (H.A.v.V.) and co-researcher (L.V.). Disagreements were resolved in consensus.
The questionnaire contained validated Dutch translations of the 12-item Short Form health survey (SF-12) to measure general health-related quality of life, the State Trait Anxiety Inventory (STAI-6) to measure generic anxiety, the Memorial Anxiety Scale for Prostate Cancer (MAX-PC), the Decisional Conflict Scale (DCS), the Center of Epidemiologic Studies Depression scale (CES-D) and the Eysenck Personality Questionnaire (EPQ) [11–16]. Details of the SF-12, STAI-6 and DCS scores, the attitude scale and PCa knowledge (15 items) have been described previously [17,18].
The MAX-PC measures PCa-specific anxiety . Two subscales were used; the PCa anxiety scale and the fear of recurrence scale, 50% of the total score (range 0–35) of both scales identifies patients who have clinically significant PCa anxiety .
Depression was assessed using the CES-D, which consists of 20 items with four response options each. Total scores range from 0 to 60. Scores of ≥16 define patients as clinically depressive .
Personality was assessed using the EPQ, which consists of 48 items with two response options each . The EPQ consists of four personality scales; psychoticism, extraversion, neuroticism and social desirability (of questionnaire response). Scale scores range from 0 to 12.
The involvement of the urologist in the decision-making process was assessed by the question ‘Who had the most influence in the treatment choice, you or your urologist?’, with five response options ‘you’ (1), ‘you/both’ (2), ‘both (3)’, ‘both/urologist’ (4), and urologist (5). We recoded these options in three decision categories: patient-based (option 1 or 2), shared (option 3) and urologist-based decision (option 4 or 5). The involvement of the environment (i.e. family, friends) was assessed through a similar question.
We assessed the differences between those who complied and those who did not comply with treatment recommendations using the chi-squared test for categorical variables and Mann–Whitney U-test for continuous variables. We used multivariable logistic regression analyses (forward likelihood ratio) to assess the influence of P(indolent), the urologist, and levels of generic anxiety on patients' compliance with AT or AS recommendations. Furthermore, we assessed the number of patients who discontinued AS and their reasons.
Analyses were performed using SPSS (version 17.0, SPSS Inc., Chicago, IL). A P value of <0.05 was considered to indicate statistical significance.
CHARACTERISTICS OF THE STUDY POPULATION
A total of 240 patients with a mean (sd) age of 64 (5) years were included. Study population characteristics are shown in Table 1. Based on the outcome of the RC, AT was recommended in 185 patients (P[indolent]≤70%) and AS in 55 patients (P[indolent] >70%, Fig. 1). Patients were compliant with RC recommendations in 176/240 cases (73%); 71% (131/185) were compliant with AT recommendations and 82% (45/55) with AS recommendations (Fig. 1).
Table 1. Characteristics and clinical characteristics of the patient population, stratified by treatment recommendation based on the outcome of the RC
|Mean age, years (sd, range)||64 (5, 55–75)||65 (5, 55–75)||0.25||66 (5, 55–75)||61 (3, 56–64)||64 (5, 55–75)|
|Marital status, n (%)|| || || || || || |
| Married or cohabiting||101 (81)||45 (85)||0.58||38 (86)||9 (90)||193 (84)|
| Single||23 (19)||8 (15)|| ||6 (14)||1 (10)||38 (16)|
|Education level, n (%)|| || || || || || |
| Low||24 (20)||10 (20)||0.08||10 (23)||1 (10)||45 (20)|
| Intermediate||57 (46)||15 (29)|| ||21 (48)||4 (40)||97 (43)|
| High||42 (34)||26 (51)|| ||13 (29)||5 (50)||86 (37)|
|Employment status, n (%)|| || || || || || |
| Paid job||43 (35)||18 (34)||0.99||10 (23)||7 (70)||78 (34)|
| Unpaid job||13 (11)||6 (11)|| ||6 (14)||0||25 (11)|
| Retired||67 (54)||29 (55)|| ||28 (63)||3 (30)||127 (55)|
|Comorbidities: median number of conditions (range)||1.0 (0–4)||1.0 (0–4)||0.49||1.0 (0–5)||0 (0–3)||1.0 (0–5)|
|Clinical characteristics|| || || || || || |
| Median PSA, ng/mL (sd, range)||8 (4.0, 1–20)||7 (3, 3–18)||0.02||5 (3, 1–17)||7 (3, 4–13)||7 (4, 1–20)|
| Mean mm cancer in biopsy (sd, range)||7 (5, 0.2–20)||3 (3, 0.1–16)||<0.001||1 (2, 0.1–11)||1 (1, 0.4–5)||5 (5, 0.1–20)|
| Mean mm healthy tissue in biopsy (sd, range)||73 (17, 40–127)||73 (19, 40–112)|| ||86 (14, 60–126)||87 (13, 66–110)||76 (17, 40–127)|
| Clinical T stage on DRE, n (%)|| || || || || || |
| T1c||76 (58)||35 (65)||0.92||31 (69)||5 (50)||147 (61)|
| T2||55 (42)||19 (35)||0.39||14 (31)||5 (50)||93 (39)|
|Mean P(indolent), %* (sd, range)||36 (17, 5–70)||55 (14, 15–70)||<0.001||81 (6, 72–97)||78 (5, 71–87)||50 (23, 5–97)|
Of the 141 patients who eventually chose AT, 103 (73%) underwent surgery, 37 (26%) underwent radiotherapy and one underwent high-intensity focused ultrasonography (<1%).
The most frequently reported advantage of AT by patients was that AT was an appropriate way to treat PCa (68/141, 48%). The side effects of AT, such as incontinence and impotence (99/141, 70%), were cited as a disadvantage by many patients (Table 2). The most frequently reported advantage of AS included the delay of any physical side effects caused by physical damage after AT, so that quality of life/lifestyle was not altered (51/99, 52%; Table 2). Quality-of-life scores were largely similar between those who complied and those who did not.
Table 2. The most reported advantages and disadvantages of AS (n= 99) and AT (n= 141) by patients. More than one answer could be given per patient
|Advantages of AS*|| |
|1. Delay of any physical side effects resulting from physical damage after AT, so that quality of life/lifestyle is not altered||51 (52)|
|2. Insight into the clinical behaviour of PCa by frequent check-ups, so buying time to think before making a treatment decision||28 (28)|
|3. Delay of (unnecessary) AT||15 (15)|
|Disadvantages of AS*|| |
|1. Uncertainty and distress about the development of the PCa||26 (26)|
|2. None||23 (23)|
|3. Risk of unfavourable consequence, such as clinical stage progression or metastases||15 (15)|
|Advantages of AT†|| |
|1. Appropriate way to treat PCa with minimum side effects, e.g. robot-assisted laparoscopic prostatectomy and radiotherapy||68 (48)|
|2. Removing the PCa||55 (39)|
|3. Certainty about healing from PCa||21 (15)|
|Disadvantages AT†|| |
|1. Side effects resulting from physical damage after AT, e.g. incontinence, impotence and bowel complaints||99 (70)|
|2. None||13 (9)|
Active treatment was recommended to 185 patients and, of these, 71% (131/185) were compliant. Of the non-compliant patients, 48% (26/54) had a P(indolent) between 60–70%. The most common reasons for urologists to be non-compliant with AT recommendations were patients' preference for AS (n= 30), patients fulfilling the inclusion criteria of the Prostate Cancer Research International: Active Surveillance (PRIAS) protocol (n= 8 [PSA ≤10 ng/mL, PSA density <0.20, clinical stage ≤ T2, Gleason sum ≤3+3 and ≤2 positive biopsy cores]) , and patients having comorbid conditions (n= 8). Patients with comorbid conditions reported that their urologists also gave other treatment options, but they preferred AS. The proportion of comorbid conditions did not differ between patients (aged 65–75 years) who chose AS or AT (P= 0.14). The most reported advantage of AS according to patients was the delay of the physical side effects of AT (28/54, 52%), and the most reported disadvantages were uncertainty and distress about the development of the PCa (15/54, 28%).
Patients who complied with AT recommendations had higher mean PSA levels (8 vs 7 ng/mL, P 0.02), a greater mean amount of cancer tissue in their biopsies (7 vs 3 mm, P < 0.001), lower mean calculated P(indolent) (36% vs 55%, P < 0.001, Table 1), higher mean levels of generic anxiety (42 vs 38, P= 0.03, Table 3), and higher mean scores on the depression scale (9 vs 8, P= 0.03) than those who did not comply. The proportion of compliant patients who were defined as clinical depressive (total scores of ≥16), however, did not differ compared with the proportion of non-compliant patients with an AT recommendation (21/120, 18% vs 10/50, 20%, P= 0.70). As expected, those who complied with AT recommendations had a positive attitude towards AT (91% vs 32%, P < 0.001), and a negative attitude towards AS (87% vs 9%, P < 0.001) more frequently than those who did not comply. The decisions of those who complied were more often patient-based than based on the urologist's opinion (30% vs 19%, Table 4) compared with those who did not comply. In multivariable analysis the strongest determinants for non-compliance were a urologist-based decision (odds ratio [OR] 5.2, 95% CI 1.5–18.6, P= 0.01), the P(indolent)(OR 1.08 per 1% increase, 95% CI 1.0–1.1, P < 0.001), and generic anxiety (OR 0.9, 95% CI 0.8–0.9, P < 0.001).
Table 3. Questionnaire scores, stratified by recommendation and compliance with recommendation
|Mean (sd) SF-12 generic health status score*|| || || || || |
| Physical Component Summary, PCS-12||52 (7)||51 (8)||0.48||51 (8)||53 (4)|
| Mental Component Summary, MCS-12||52 (10)||54 (10)||0.09||53 (10)||52 (12)|
|Mean (sd) STAI-6 generic anxiety score†||42 (10)||38 (10)||0.03||39 (10)||41 (16)|
|Mean (sd) MAX-PC score|| || || || || |
| Subscale PCa Anxiety§||10 (7)||12 (9)||0.31||13 (8)||11 (8)|
| Subscale Fear of Recurrence‡||7 (2)||8 (3)||0.12||7 (2)||7 (3)|
| Total of both subscales||17 (6)||19 (7)||0.05||20 (7)||18 (7)|
|Mean (sd) Decision Conflict Scale score¶||27 (13)||26 (15)||0.80||28 (15)||25 (12)|
|Mean (sd) CES-D score††||9 (8)||8 (10)||0.03||8 (7)||9 (10)|
|Mean (sd) EPQ four personality scales score‡‡|| || || || || |
| Psychoticism||3 (2)||3 (1)||0.82||3 (1)||2 (1)|
| Extraversion||8 (3)||7 (3)||0.14||7 (3)||7 (3)|
| Neuroticism||4 (3)||3 (3)||0.38||3 (3)||3 (3)|
| Social desirability||8 (3)||8 (2)||0.54||8 (3)||8 (3)|
Table 4. Knowledge scores, attitude, and the influence of the urologist and the environment on patients in making their treatment decision
|Mean PCa knowledge* (sd, range)||9 (2, 2–13)||9 (2, 3–12)||0.34||9 (1, 5–11)||9 (2, 5–12)|
|Attitude towards AT, n (%)|| || || || || |
| Negative attitude||10 (9)||32 (68)||<0.001||19 (58)||0|
| Positive attitude||98 (91)||15 (32)|| ||14 (42)||9 (100)|
|Attitude towards AS, n (%)|| || || || || |
| Negative attitude||91 (87)||4 (9)||<0.001||8 (24)||8 (89)|
| Positive attitude||14 (13)||43 (91)|| ||26 (76)||1 (11)|
|‘Who has the most influence in the treatment choice, the patient or the urologist?’, n (%)|
| Patient||37 (30)||10 (19)||0.11||8 (18)||7 (70)|
| Shared decision||66 (54)||28 (53)|| ||28 (64)||3 (30)|
| Urologist||20 (16)||15 (28)|| ||8 (18)||0|
|‘Who has the most influence in the treatment choice, the patient or his environment?’, n (%)|
| Patient||77 (63)||38 (72)||0.44||31 (70)||8 (80)|
| Shared decision||45 (37)||15 (28)|| ||11 (25)||2 (20)|
| Environment||1 (1)||0|| ||2 (5)||0|
Ten of 55 patients were non-compliant with an AS recommendation, resulting in limited reliability for comparisons with compliant patients (Tables 1, 3 and 4), therefore, no statistical testing was done. The most common reason for urologists to be non-compliant was that patients wanted AT (n= 7). Reasons for patients' non-compliance included anxiety about progression of the PCa, too much stress involved in AS and undergrading of the PCa. Most patients reported removal of the PCa as the advantage of AT (n= 4).
FOLLOW-UP OF AS PATIENTS
In the present study, 99 patients initially chose AS; 11% (11/99) were lost to follow-up and 14% (14/99) discontinued AS. The mean (range) follow-up of the 74 patients on AS was 12 (0–26) months. AS was discontinued because patients wanted AT (4/14, mean follow-up 6 months) or because of PCa progression (10/14, mean follow-up 15 months). The proportion of patients who discontinued AS below or above the 70% threshold did not differ: 16% (8/50) and 16% (6/38), respectively (P= 0.98). Reasons for discontinuing AS did not differ between either group (P= 0.73).
In the present study, where the RC was actively implemented into clinical practice, urologists and patients were compliant with AS recommendations based on the RC in most cases (45/55, 82%), but AS was chosen in 54 of 185 cases (29%) where AT was recommended. These patients had relatively high calculated P(indolent), lower levels of generic anxiety, and the influence of the urologist in treatment decision-making was stronger compared with that in patients who were compliant with AT recommendations. The most common reason for urologists to opt for AS instead of AT was that patients preferred AS. This indicates that the threshold for AS of >70% may be too high for many patients. This form of non-compliance may also be explained by the fact that urologists had a preference for AS, particularly in patients who fulfilled the inclusion criteria of the PRIAS protocol (59%, 32/54, P[indolent] range 23–70%) . Their relatively low P(indolent) was caused by a higher mean mm cancer/mm benign tissue ratio (4.1% vs 1.1%, P < 0.001) at biopsy and a higher mean PSA density (0.15 vs 0.11, P < 0.001) than in patients with a P(indolent) >70% and who fulfilled the PRIAS inclusion criteria (76%, 34/45). The proportion of patients who discontinued AS and their reasons for discontinuing AS, i.e. patients preferred AT or had PCa progression below or above the 70% threshold, did not differ. A reason reported by urologists for some patients' non-compliance with the AT recommendations based on the RC, was that these patients preferred AS. It may be possible that urologists had recommended AT, but patients were not willing to undergo AT. In those cases urologists could not be denoted as being non-compliant with the AT recommendations based on the RC. Conversely, urologists gave an AT recommendation in some patients with P(indolent) >70%. Ultimately, it remains the patient's decision to accept or decline the AS or AT recommendation based on the RC, reflecting a personal threshold for the probability of having potentially indolent PCa.
The non-compliance of patients could be influenced by the treatment preferences of urologists, the way urologists communicate treatment options with their advantages and disadvantages, impact on quality of life, and patient's calculated P(indolent). Patients may also be influenced by information from other sources, e.g. the Internet, leaflets, family, friends and second opinions [22–25]. Patients who experience higher levels of generic anxiety may opt for AT rather than AS, because they have difficulties with living with untreated PCa and/or have more anxiety about PCa progression. In the present study, those who complied with AT recommendations had higher levels of generic anxiety than those who did not. This is in contrast to a previous study where anxiety has not been shown to be higher in men who have chosen initial treatment versus AS . The present study confirmed that the most common reported advantage of choosing AS for those who did not comply with AT recommendations was the delay of physical side effects after AT, so that quality of life/lifestyle was not altered [18,24].
The present study is one of the first to investigate urologists' and patients' compliance with recommendations based on a RC. Nomograms have a long track record in urology, and studying their impact on clinical practice is important. Evaluation of impact requires setting a threshold to recommend AS vs AT . This threshold can be defined by a careful weighing of the risks and benefits in a full decision analysis . In the present study, the 70% probability threshold was primarily motivated by a high sensitivity to actively treat potentially important PCa in a screening setting. We did not correct the probability threshold and calculated P(indolent)s for use in a clinical setting. Since the introduction of the PSA test, a favourable stage shift at the time of detection has been observed. The proportion of T1c cancers at the initial screening round of the ERSPC in Rotterdam was 47.8% during the years 1994–1998, while in the control arm, reflecting the clinical setting, this proportion was 28.5% and increased during the years 2003–2006 to 50% . In the present study the proportion of T1c tumours (61%, Table 1) bears more similarity with the screening setting in the period 1994–1998 from which the RC is derived. This increase in the proportion of T1c cancers reflects the increase of PSA testing in the clinical setting.
The results of the present implementation study showed that the RC may well be of use in treatment decision-making. The P(indolent) threshold of >70% may be suitable for AS strategies. Urologists used the RC in most eligible patients diagnosed with PCa (93%, 67/72) for whom RC level three was used previously; however, we do not know whether urologists will continue to use the RC after the project. The best predictors of whether physicians will use a prediction rule are acquired familiarity, confidence in the usefulness of the rule and its user-friendliness . Urologists may be more used to the inclusion criteria of an AS protocol, such as the PRIAS protocol, than the use of a RC to select men for AS or AT; however, the RC does not only support the selection of patients for AS or AT (as a decision tool), but also informs urologists and patients about the probability of a potentially indolent PCa (as a nomogram for a personal decision threshold for the risk of a potentially aggressive PCa).
Limitations of the study are that it is not clear how the motives of patients in choosing AS or AT developed, especially the patients who chose AS against the RC recommendations, and that it is not clear how the outcome of the RC affects patient's choice and the urologist in his/her counselling. Further research is needed into these topics, and a longer follow-up of patients on AS is important to improve and validate the chosen 70% threshold for indolent disease. This threshold or lower appeared to be acceptable in this Dutch clinical cohort but may not be acceptable elsewhere, reflecting factors such as cultural differences.
In conclusion, AS recommendations were followed by most patients, while 29% of patients with AT recommendations chose AS. Although further research is needed to improve the probability threshold for recommending AS over AT, the current RC proved to be useful in treatment decision-making in patients with localized PCa.
We thank the nurses/clinical nurse specialists for supporting the urologists in the use of the RC, the data collection and data processing: A. Verkerk (Erasmus Medical Center Rotterdam), F. van Renen-Bolier and J. Tillema (University Medical Center Groningen), W. de Blok (Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital Amsterdam), and J. Vekemans (Elisabeth Hospital Tilburg). We also thank I. van Onna and P. van den Broeke (urologists at Amphia Hospital), and I. de Jong (urologist at UMCG) for participating in this study. Finally, we want to thank L. Venderbos for grouping and scoring the advantages and disadvantage of patient's treatment choices.
This work was financially supported by grants of the Dutch Organization for Health Research and Development (ZonMw) and the Prostate Cancer Research Foundation (SWOP) Rotterdam, The Netherlands.
CONFLICT OF INTEREST