Choice between prostatectomy and radiotherapy when men are eligible for both: a randomized controlled trial of usual care vs decision aid


  • Trial NTR1334

Correspondence: Julia J. van Tol-Geerdink, Department of Radiation Oncology (874), Radboud University Nijmegen Medical Center, PO Box 9101, 6500 HB Nijmegen, The Netherlands.



What's known on the subject? and What does the study add?

  • Many patients are eligible for more than one treatment option for prostate cancer. In usual care, urologists have a large influence on the treatment choice. Decision aids, providing balanced information on the pros and cons of different treatment options, improve the match between patient preferences and treatment received.
  • In men eligible for both surgery and external beam radiotherapy, treatment choice differed by hospital. Across the participating hospitals, the decision aid consistently led to fewer patients remaining undecided on their treatment preference and more patients choosing brachytherapy.


  • To examine the treatment choice for localized prostate cancer in selected men who were eligible for both prostatectomy and radiotherapy.
  • To examine whether increased patient participation, using a decision aid, affected the treatment choice.

Patients and Methods

  • From 2008 to 2011, 240 patients with localized prostate cancer were enrolled from three separate hospitals.
  • They were selected to be eligible for both prostatectomy and external beam radiotherapy. Brachytherapy was a third option for about half of the patients.
  • In this randomized controlled trial, patients were randomized to a group which only discussed their treatment with their specialist (usual care group) and a group which received additional information from a decision aid presented by a researcher (decision aid group). The decision aid was based on a literature review.
  • Predictors of treatment choice were examined.


  • Treatment choice was affected by the decision aid (P = 0.03) and by the hospital of intake (P < 0.001).
  • The decision aid led to more patients choosing brachytherapy (P = 0.02) and fewer patients remaining undecided (P < 0.05).
  • Prostatectomy remained the most frequently preferred treatment.
  • Age, tumour characteristics or pretreatment urinary, bowel or erectile functioning did not affect the choice in this selected group.
  • Patients choosing brachytherapy assigned more weight to convenience of the procedure and to maintaining erectile function.


  • Traditionally, patient characteristics differ between surgery and radiotherapy groups, but not in this selected group of patients.
  • Men eligible for both prostatectomy and radiotherapy mostly preferred prostatectomy, and the treatment choice was influenced by the hospital they visited.
  • Giving patients evidence-based information, by means of a decision aid, led to an increase in brachytherapy.


For primary localized prostate tumours, different treatment options are available. The most frequently applied treatments are radical prostatectomy and radiotherapy, either external or interstitial (brachytherapy). These treatments can generally offer comparable results in terms of tumour control but differ in their profile of side effects. Prostatectomy is associated with a higher risk of urinary incontinence and erectile dysfunction, whereas radiotherapy is more likely to cause persistent bowel problems. In this choice, patients often rely on the advice of the specialist [1, 2]. However, the recommendations of specialists tend to be influenced by their specialty. Most urologists recommend prostatectomy, whereas the majority of radiation oncologists consider the different treatments to be equivalent [3, 4].

Current guidelines [5, 6] no longer indicate a single treatment as the optimal treatment of localized prostate carcinoma. Therefore patients should be involved in the treatment decision, which calls for the use of decision aids [7]. A decision aid provides structured and balanced information on the pros and cons of different treatment options. Decision aids increase patient's knowledge, help to clarify preferences and reduce uncertainty about the treatment decision [8, 9].

This randomized controlled trial examined the effect of a decision aid on the treatment choice for prostate cancer. The decision aid provided information on radical prostatectomy, brachytherapy by means of seed implants and external beam radiotherapy. The decision aid included information on the procedures, the likelihood of cure and side effects in the urinary, bowel and sexual domain for the three different treatments. The treatment choice with the decision aid was compared with that in patients receiving usual care.

For this study, we selected patients to be eligible for prostatectomy as well as radiotherapy. In most previous studies comparing prostate cancer treatments, patients' characteristics differed. For example, surgery patients were often younger and had less advanced tumours then irradiated patients [10-12]. By selecting, this study aimed to involve a more homogeneous population that actually had a choice. The aim of this study was to examine the effect of a decision aid on the treatment choice for localized prostate cancer in men who really have a choice.

Patients and Methods


Between March 2008 and February 2011, patients with primary localized prostate cancer (T1–3aN0M0), intending to be treated and eligible for both radiotherapy and radical prostatectomy, were asked to participate in the study. Exclusion criteria were contra-indications for surgery (based on for example age or cardiovascular problems) or external radiotherapy (based on for example Crohn's disease), mental or cognitive problems as assessed by the physician, inadequate knowledge of the Dutch language or a preference for active surveillance. We excluded active surveillance patients because our decision aid did not include risk information on this option.

Brachytherapy was offered only to a selected group of patients. Exclusion criteria for brachytherapy were a small or large prostate volume (<20 mL or >50 mL), PSA > 15, Gleason >7 and/or severe urinary symptoms (requiring medication or, if available, IPSS > 12 and/or Qmax <15 mL/s).

Recruitment took place in three locations in the Netherlands, i.e. the Radboud University Nijmegen Medical Centre, the Canisius Wilhelmina Hospital in Nijmegen and the Rijnstate Hospital in Arnhem. The last two are large non-academic centres. The study was approved by the research ethics committees of these hospitals.


Patients were monitored in a prospective randomized controlled trial between usual care and a decision aid (Fig. 1). Patients with positive biopsies were informed by their urologist about the fact that different treatments were available. The urologists were instructed to mention the treatment options briefly and not to reach a treatment decision in the first consultation. The study was mentioned by the urologist and informed consent was obtained after the patient was informed about the study by the researcher. A second consultation with the urologist was scheduled to elaborate on the treatment choice.

Figure 1.

Study design. Patients were randomized in a 2 : 1 ratio to the decision aid group and the usual care group.

Enrolled patients were randomized within each hospital to (i) the usual care group, discussing the treatment choice with their specialist, or (ii) the decision aid group, also discussing the treatment choice with their specialist but after a consultation with the researcher who presented a decision aid. Patients were assigned in a 1:2 ratio to the usual care group and the decision aid group, respectively, to have a decision aid group that was large enough to answer other research questions, reported elsewhere. Patients in both randomization arms filled out questionnaires about their health and their opinion on the treatment options.

Presentation of the decision aid

The decision aid was delivered by the researcher in a semi-structured interview. The main features of the treatments were described, including the treatment process and the salvage options. This was done for radical prostatectomy (by open, laparoscopic or robot-assisted procedure), external beam radiotherapy and, if applicable, brachytherapy. Thus, according to whether men were eligible for brachytherapy, as described in the Patients section, they were offered two (radical prostatectomy and external beam radiotherapy) or three (radical prostatectomy, brachytherapy and external beam radiotherapy) treatment options.

Once all procedural and outcome information was presented, patients were asked which treatment they preferred. They were allowed some time to consider their preference and received the information in writing to take home, including a description of the treatment options and the numerical information terms of cure and side effects as in Fig. 2. Their preference was assessed – or confirmed – by telephone after 3 days. If patients indicated a preference, the preferred treatment was documented in the medical records.

Figure 2.

Decision aid for patients eligible for all three treatments.

Development of the Decision Aid

A literature search was carried out to obtain risk information on the probabilities of progression, survival and side effects (urinary, bowel and erectile). The decision aid did not include the option of active surveillance because insufficient risk information was available in the literature to compare it with the other treatment options.

Literature on tumour control

Information was gathered on the likelihood of 5-year disease-free survival and the 10-year disease-specific mortality. These outcomes take regular practice into account, including salvage therapy and treatment-induced mortality. No randomized trials were available comparing the effects of the treatments. From the available studies [13-20], weighted means were calculated, based on the sample size in each study. Separate estimates were calculated for low/intermediate risk (T1T2 and Gleason ≤7 and PSA ≤20) and high risk patients.

Literature on side effects

The main side effects were identified as erectile dysfunction and late bowel and urinary problems. Obtaining numerical information on morbidity was a challenge. No randomized controlled trials, randomizing between prostatectomy, external beam radiotherapy and/or brachytherapy, were available. The likelihood of side effects varied widely in the literature due to differences in symptom definitions, patient characteristics, pretreatment functioning and follow-up duration. To tackle differences in any other symptom definitions, we selected studies in which two or more treatment options were compared on side effects. In this way we could ascertain that the same definition was used for the side effects when comparing different treatments. This yielded 18 studies [21-38]. Data were obtained for genito-urinary morbidity (incontinence requiring two pads or more per day), gastrointestinal morbidity (diarrhoea) and erectile dysfunction (erections insufficient for intercourse). If available, data were selected for patients under 70 years of age to reduce differences based on age disparities between patients having surgery or radiotherapy. Outcomes appeared to be influenced by baseline functioning. For example, the likelihood of erectile dysfunction after treatment is strongly affected by pretreatment potency. Therefore, if possible, the risk for each side effect was calculated correcting for baseline functioning. Furthermore, data were gathered on the side effects after a follow-up duration of 24 months.

These analyses formed the basis for a consensus discussion with a dozen specialists (both urologists and radiation oncologists). After minor adjustments, consensus was reached and the resulting numbers were imported in the decision aid. The numbers were presented by means of pie charts (Fig. 2).


Data were collected on several variables that were expected to affect treatment choice. The assessments were at baseline (t1), i.e. before the treatment options were discussed, and at pretreatment (t2), i.e. after the treatment was chosen but before it was carried out.

Patients' demographic and medical characteristics

Self-report data were collected on demographic variables (age, marital status, having children, working status, education and religion). Tumour characteristics (T status, pretreatment PSA value and Gleason score) were extracted from the medical records. Baseline functioning was assessed by questionnaire as prostate-specific health-related quality of life by means of the expanded prostate cancer index composite (EPIC) [39].

Treatment preference and treatment received

At baseline (t1), patients were asked by questionnaire if they already had a baseline treatment preference. After the second consultation with the urologist, when the treatment choice was made, data were gathered by questionnaire (t2) on their final treatment preference. The time between diagnosis and decision was 2 weeks on average (14 ± 6 days).

Patients assigned an importance weight on a five-point scale at t2, ranging from ‘not at all important’ to ‘very important’, for outcomes such as the probability of dying from prostate cancer, the risk of urinary problems, of bowel problems and of sexual problems, and the burden of the procedure itself (surgery or radiotherapy). The treatment received was extracted from the medical records.


For patients' preferences and treatments received (prostatectomy, external beam radiotherapy or brachytherapy), multinomial univariate analyses were conducted using chi-squared tests. The tests were done separately for the final treatment preference and treatment received, comparing the groups with regard to age, demographic variables, tumour characteristics, health-related quality of life, the recruiting hospital and the use of the decision aid. For these analyses continuous variables were dichotomized (split at the median value). Multivariable analyses were conducted using multinomial logistic regression on those variables that were significant in univariate analyses (P < 0.05).

The agreement between final treatment preference and treatment received was assessed with κ. The reasons for a treatment preference were analysed by comparing the weights assigned to different outcomes on a five-point scale. Paired t tests were used to test whether importance weights differed.



In total, 307 patients were asked to participate in the study. Of these patients, 36 declined (12%) and others were excluded because additional screening revealed other health problems (n = 14) or because they chose active surveillance after all (n = 17). This yielded 240 patients giving informed consent for this study. The eligibility for brachytherapy was assessed for each patient in the decision aid group. Most patients (56%) were eligible for all three treatment options, i.e. prostatectomy, brachytherapy and external beam radiotherapy, whereas 44% were eligible for two options, i.e. prostatectomy and external beam radiotherapy.

The patients in the decision aid group were comparable to their counterparts in the usual care group for all tumour-related and other characteristics (Table 1). Between hospitals, differences were only found in education (P = 0.02), with more highly educated patients in clinic A, and in T status (P < 0.01), with more patients having a low risk T status in clinic B.

Table 1. Patients in the usual care group (n = 77) and the decision aid group (n = 163)
 Usual care (%)Decision aid (%)
  1. *In 4 weeks prior to intake. None of the differences was statistically significant.
Education: college or higher3232
Living with partner9386
Age, years (mean [SD])64 (5)64 (5)
Baseline functioning*  
No urinary pads required9798
No loose stool5758
Erections firm enough for intercourse5648
Tumour and hospital  
Gleason ≤79597
PSA ≤208996
Accrual in hospital A2729
Accrual in hospital B5249
Accrual in hospital C2122
Baseline treatment preference  
Radical prostatectomy4946
External beam radiotherapy76
No preference indicated3236

Treatment Overview

Overall, the most frequently received treatment was radical prostatectomy (71%), which was also indicated by most patients (67%) as their final treatment preference (Table 2). The prostatectomies were performed using open surgery (15%), laparoscopic (25%) or robot-assisted laparoscopic (61%) procedures.

Table 2. Patients' final treatment preferences and treatments received in the usual care group (n = 77) and the decision aid group (n = 163)
 RP (%)BT (%)EBRT (%)Undecided (%)
  1. RP, radical prostatectomy; BT, brachytherapy; EBRT, external beam radiotherapy.
Treatment preference6717134
Treatment received711218

The majority of patients received their preferred treatment (κ = 0.85). Patients who indicated no preference received either prostatectomy (n = 3) or external beam radiotherapy (n = 5).

Decision Aid

When comparing the decision aid group with the usual care group (Table 3), an effect on final treatment preference was found (P = 0.03). The decision aid group preferred brachytherapy more frequently (20% vs 8%, P = 0.02) and remained undecided less frequently (2% vs 8%, P < 0.05) than the usual care group. The treatments received by the patients were also affected by the decision aid (P = 0.04).

Table 3. Effect of the decision aid on final treatment preferences and treatments received in the usual care group (n = 77) and the decision aid group (n = 163)
 RP (%)BT (%)EBRT (%)Undecided (%)P
  1. RP, radical prostatectomy; BT, brachytherapy; EBRT, external beam radiotherapy.
Treatment preference    0.03
Usual care group738128
Decision aid group6520132
Treatment received    0.04
Usual care group78418
Decision aid group681517


The hospital of intake also appeared to have an effect on treatment (P ≤ 0.001). Multinomial logistic regression analysis showed that the final treatment preferences and the treatments received were affected by both the decision aid (P < 0.05) and the hospital (P < 0.001). No interaction between decision aid and hospital was found.

Patient Characteristics and Functioning

The treatment received was not associated with the patients' age, demographic variables, tumour characteristicsor function scores. In other words, these variables did not predict treatment received.

Reasons for Preferences

All patients were asked to assign an importance weight to several outcomes in the questionnaire at t2 (Table 4). Overall, survival was rated equally important to bowel or urinary side effects for the treatment decision (4.2, 4.1, 4.0 respectively). Sexual problems (3.3) and the inconvenience of the procedure itself (3.5) were generally rated as less important (P < 0.001). However, analysing the patient groups by final treatment preference revealed remarkable differences (Table 4). Patients preferring brachytherapy considered sexual problems and the burden of the procedure more often important (P < 0.05 and P < 0.01 respectively) than other patients.

Table 4. Importance of specific outcomes for the final treatment preference: percentages of patients indicating the outcome to be ‘important’ or ‘very important’ for their preference
 RP (N = 154) (%)BT (N = 25) (%)EBRT (N = 35) (%)
  1. *P < 0.05, **P < 0.01 difference between the patients preferring brachytherapy and the patients preferring radical prostatectomy or external beam radiotherapy. RP, radical prostatectomy; BT, brachytherapy; EBRT, external beam radiotherapy.
Urinary problems759278
Bowel problems808874
Sexual problems4168*49
Convenience of procedure5384**66


This study showed that treatment choice was affected by the use of a decision aid and by the hospital. Age, demographic variables, baseline functioning scores and tumour characteristics did not differ between patients receiving surgery or radiotherapy.

Decision Aid

While baseline preferences were similar between both study arms (Table 1), the decision aid affected the final treatment choice. Prostatectomy was chosen less frequently in the decision aid group than in the usual care group (68% vs 78% respectively). Auvinen and coworkers [40] found that, after patients were informed with a decision aid, prostatectomy was chosen less frequently and external beam radiotherapy more frequently (no brachytherapy was offered). In our study, the use of the decision aid also appeared to shift the patients' preference somewhat from prostatectomy to radiotherapy, be it brachytherapy. Possibly, the information that cure was comparable after all treatments was responsible for this shift; when the information was presented, several patients spontaneously mentioned that they had assumed that cure after prostatectomy would be much better than after radiotherapy. Our results are in line with previous research showing that decision aids may cause patients to choose less intensive treatments [9, 41].

Differences between Hospitals

Variation in treatments between the three participating hospitals was found. This may be related to differences between clinics and between patient populations, e.g. more high risk patients. A strong point of this study, however, is that despite these differences the effect of the decision aid remained significant when taking the differences between hospitals into account and no interaction between hospital and decision aid was found. In other words, the decision aid affected treatment choice in a similar way in all participating hospitals.

Patient and Tumour Characteristics

In this study, age, demographic variables and tumour characteristics were not related to the treatment received. In most previous studies comparing prostate cancer treatments, patient groups differed with regard to patient and tumour characteristics. For example, in a study on the choice between prostatectomy and brachytherapy, age and tumour grade were found to be strong predictors of patient preferences [42]. In that study, however, several patients could not choose brachytherapy because of a high Gleason score, or conversely could not choose prostatectomy because they were too old (age range up to 89 years). In our study, in contrast, patients were selected to be eligible for both surgery and radiotherapy. Thus, patients were more comparable at baseline and actually had a choice. In this selected patient group, age and tumour characteristics did not appear to determine treatment choice. To our knowledge this is the first study in which patients receiving surgery were comparable to those receiving radiotherapy with regard to patient and tumour characteristics.

Reasons for Preference

Patients favouring brachytherapy assigned a relatively higher importance to the convenience of the procedure, in line with previous reports [43], and to the risk of sexual problems. In the literature, impotence is usually reported to be less important than urinary incontinence to patients with prostate cancer [44]. Our results imply, however, that this appears to be different for the group of patients choosing brachytherapy.


Based on the results of this study, we recommend the use of decision aids, especially in situations where different treatments are available with comparable likelihood of cure. The aim of decision aids is to help patients reach decisions. In this study, the decision aid succeeded in leaving fewer patients undecided.

We also recommend that future studies, aiming to compare the effects of different treatments for localized prostate cancer, should be carried out on patients that are selected to be eligible for the different treatment options. This selection results in a more homogeneous patient population, which enables a better comparison of the treatment effects.


The limitations of this study are the relatively small number of patients in the irradiated groups. In addition, one treatment option, active surveillance, was not included in the decision aid. This study therefore focused on patients intending to choose one of the more active treatments.


By selecting patients who were eligible for both surgery and radiotherapy, we studied patients who were more similar at baseline, compared with previous studies, and who actually had a choice. Under these circumstances, the treatment received was not determined by age and tumour characteristics but by other factors, such as the (in)convenience of the procedure. The decision aid had a significant impact on treatment choice by increasing the number of patients choosing brachytherapy and reducing the number of patients remaining undecided.

Financial Support

Financial support for this study was provided by a grant (2007-3809) from the Dutch Cancer Society, Amsterdam, The Netherlands.

Conflict of Interest

None declared.