Institutional availability, training resources, and supervised participation in robot-assisted surgery among Polish urology residents: a nationwide survey

Abstract

Robot-assisted surgery (RAS) has expanded rapidly in Poland, but its integration into resident education remains uncertain. In a nationwide anonymous 35-question survey sent to 390 Polish urology residents, 156 completed it (40.0%). Among respondents, 110 (70.5%) reported RAS availability in their training centre; 56 (35.9%) reported robotic training access, 43 (27.6%) a robotic course, 46 (29.5%) simulator availability, 76 (48.7%) assistant participation, and 16 (10.3%) a supervised robotic procedure or procedural step. All respondents with training access were in RAS centres. Among these 110 respondents, course attendance (aOR 5.38, 95% CI 2.05–14.14) and simulator availability (aOR 3.90, 95% CI 1.54–9.89) were associated with access. Training access was associated with assistant participation (aOR 6.66, 95% CI 3.06–14.50), higher current and expected self-rated skill (common aORs 12.30 and 12.15), and at least satisfactory training satisfaction (aOR 17.01, 95% CI 6.35–45.57). All 16 such reports occurred among residents with access; in an exploratory conditional Firth model, each additional residency year was associated with this outcome (aOR 2.11, 95% CI 1.30–3.97). Items were independent, not sequential stages. Institutional adoption exceeded reported training access and supervised participation. Broad survey content makes participation driven solely by robotics interest less likely, but the 40.0% completion proportion, self-report, unmeasured trainee selection, and rare supervised outcome limit inference.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11701-026-04014-z.

Introduction

Robot-assisted surgery (RAS) is now widely used in urology, but access to training remains uneven. A robotic system in the hospital does not by itself provide residents with structured instruction or supervised operating experience. A 2024 systematic review found that contemporary curricula commonly combine simulation, structured clinical exposure, and assessment against defined performance standards, although evidence of their effect on clinical outcomes remains limited [1]. An international Delphi consensus likewise identified common curriculum components and proficiency-based progression as key principles for programme design [2].

International surveys show substantial heterogeneity in robotic training. In the ESRU-ESU-YAU UROTECH survey, direct involvement in patient procedures was reported by 25.0% of respondents for robotic surgery and 58.8% for laparoscopy, and fewer than one-quarter described a standardised pathway [3]. In the Netherlands, robots were available in 75.7% of represented hospitals, but only 14.2% of residents had operated as a console surgeon [4]. In Spain, 52% had no robotic system; among those with access, 25% reported theoretical training, 23% practical training, and 13% a formal programme [5]. US programmes varied in mandatory simulation and curriculum structure [6]. The Junior-ERUS/YAU survey reported simulators in 89% of institutions and dual consoles in 65%, but novice trainees most often participated in only about 20% of robotic cases [7].

Poland provides a high-adoption setting in which to examine educational integration of RAS. Our previous national survey documented limitations in laparoscopic simulation and advanced operative exposure during urology residency [8]. In 2025, 93 hospitals performed 24,585 robot-assisted operations, including 15,032 urological procedures (61% of the national robotic volume) [9]. The national curriculum for physicians entering urology training from spring 2023 includes a 2-week robotic rotation covering safety, qualification, docking and assisting, simulator work when available, and video review, but no numeric minimum for console cases or modular console steps [10]. Because all residency years were surveyed, some respondents may have trained under earlier programme versions; the curriculum is contextual rather than an exposure shared by all participants.

The need for structured robotic curricula, competency-based progression, and appropriate candidate selection is established. This study instead quantifies how institutional uptake, educational resources, self-reported access, and clinical participation coexist in a rapidly expanding national system. We described the prevalence of institutional RAS availability, course attendance, simulator availability, self-reported access to any robotic training, assistant participation, and a supervised robotic procedure or procedural step, together with their associations. Secondary aims addressed self-rated skill, satisfaction, barriers, and procedural preferences. The survey was not designed to assess curriculum effectiveness or compliance, fairness of trainee selection, sequential progression, or objective competence.

Materials and methods

Study design and survey administration

This nationwide, cross-sectional, anonymous web-based survey assessed surgical education among physicians undergoing urology specialty training in Poland, with the present report focused on RAS-related items. The accessible sampling frame comprised 390 residents listed in the Polish Urological Association mailing database. Invitations were sent on 18 December 2025, followed by one reminder on 12 February 2026; data collection closed on 2 March 2026. Because the entire accessible population was invited, no formal sample-size calculation was performed.

Personalised links restricted participation to invited residents and prevented duplicate submissions. Link metadata were not retained with the de-identified responses. We could therefore not determine how many residents opened the survey or compare respondents with non-respondents. The export contained 195 initiated questionnaires; 39 ended before the final survey page and were excluded, leaving 156 complete questionnaires. The invitation-based completion proportion used all 390 invited residents as the denominator. It describes participation in the accessible national sampling frame and does not establish representativeness of all Polish urology residents.

Questionnaire and variable definitions

The 2026 instrument comprised 35 numbered question groups, several containing subitems, and was presented as a broad evaluation of laparoscopic and robotic training and urology residency quality rather than a robotics-only survey. It covered resident and centre characteristics; overall training quality, operative access, mentoring and external education; perceived procedural and knowledge gaps; programme preparedness and support; EAU/EBU topics; laparoscopic and robotic infrastructure, volumes, courses, simulation and practical participation; Urology BootCamp and E-BLUS; procedural preferences; self-rated skills; modality-specific satisfaction; specialty choice; and broader determinants of training quality. It was based on our earlier national survey [8] and expanded to include RAS. This article analyses the RAS-related items; the wider laparoscopic and general residency findings are outside its scope. Core definitions appear in Online Resource 1, Table S4, and the complete English translation of the administered instrument in Online Resource 2.

The questionnaire was a descriptive educational-needs assessment rather than a psychometric scale; no composite score was calculated, and the instrument was not formally validated as a scale. Training-centre type was reported as clinical, regional/voivodeship, district/county, or other. For parsimonious adjusted models, the latter three categories were grouped as non-clinical. Geographic region and centre identifiers were not collected, precluding centre-level clustering or category-specific completion proportions. Residency year was recorded from 1 to 6; years 1–3 were classified as junior and years 4–6 as senior.

Robotic course attendance meant participation in at least one national or international course. Reported institutional RAS availability, self-reported access to any robotic training, simulator availability, assistant participation, and reported performance of a supervised robotic procedure or procedural step were analysed as binary variables. The general access item did not require a defined curriculum, minimum duration, location, or number of sessions. Course attendance and simulator availability were separate, non-exclusive items and could overlap conceptually with the general access item. These measures were treated as independent indicators and not as sequential stages. Assistant participation did not record case numbers or specific bedside tasks.

A supervised robotic procedure or procedural step was recorded when a resident selected the response option “Performing a robotic procedure or procedural step under specialist supervision.” The item did not identify whether the report concerned a complete operation or an individual procedural step, record console time, quantify repeated exposure, or assess independence. This outcome was therefore not treated as evidence of independent robotic competence. The survey also did not ask whether respondents had sought robotic training, met local eligibility criteria, underwent formal selection, were offered a training position, or were limited by programme capacity.

Current and expected end-of-training robotic skill were rated in six ordered categories: 0 (none), 1 (very poor), 2 (poor), 3 (satisfactory), 4 (good), and 5 (very good). Satisfaction was rated from 0 (very dissatisfied) to 4 (very satisfied). For descriptive reporting, scores of at least 3 for skill and at least 2 for satisfaction were grouped as at least satisfactory. Full observed skill categories were retained in proportional-odds models. Because upper satisfaction categories were sparse, satisfaction was dichotomised as below satisfactory versus at least satisfactory. All ratings were self-reported and were not interpreted as objective assessments of competence or programme quality.

Data quality and outcome definition

The row-level response export was checked against the questionnaire, and all descriptive and multivariable analyses were performed from the same 156-record dataset. Values outside prespecified response options were treated as missing, and no values were imputed. Named platform and annual procedure volume were not included in inferential models because they were centre-level self-reports and centre identifiers were unavailable. The principal outcomes were self-reported access to any robotic training, assistant participation, reported performance of a supervised robotic procedure or procedural step, current and expected self-rated skill, and robotic training satisfaction.

Statistical analysis

Analyses were performed in Python 3.13.5 using NumPy 2.3.5, SciPy 1.17.0, statsmodels 0.14.6, and scikit-learn 1.8.0. Categorical variables are reported as n/N (%) and ordinal variables as median [interquartile range (IQR)]. Junior-versus-senior and clinical-versus-non-clinical comparisons used Pearson chi-square or Fisher exact tests, as appropriate. Paired current and expected skill scores were compared with the Wilcoxon signed-rank test. All tests were two-sided; no multiplicity adjustment was applied.

Covariates were selected a priori on educational and institutional grounds; no automated or significance-based selection was used. Residency year represented cumulative opportunity for exposure, centre type differences in educational environment and specialised resources, and sex a potential source of differences in opportunity or self-assessment. Because no respondent without institutional RAS availability reported training access, the access model was restricted to the 110 respondents in RAS centres to avoid complete separation; predictors were residency year, centre type, sex, robotic course attendance, and simulator availability. Assistant participation and satisfaction were modelled in all 156 respondents using residency year, centre type, sex, and training access. Current and expected skill used proportional-odds ordinal logistic regression with the same covariates.

Ethics and reporting

Participation was voluntary, and submission of the questionnaire indicated informed consent to participate. No patient data or directly identifiable personal data were collected. Under applicable Polish regulations and institutional policy, formal ethics committee review was not required for this anonymous educational survey. Consent for publication was not applicable because only aggregate, non-identifiable data are reported. Survey reporting followed the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) [11] and the Consensus-Based Checklist for Reporting of Survey Studies (CROSS) [12].

Results

Study population and robotic training exposure

The analysis included 156 complete questionnaires, corresponding to 40.0% of the 390 invited residents. The distribution by residency year was 7 (4.5%), 18 (11.5%), 29 (18.6%), 33 (21.2%), 29 (18.6%), and 40 (25.6%) for years 1 through 6; 102 (65.4%) were senior residents, and 134 (85.9%) were men. Centre categories were clinical in 51 (32.7%), regional/voivodeship in 56 (35.9%), district/county in 27 (17.3%), and other in 22 (14.1%). RAS availability was reported by 110 (70.5%), training access by 56 (35.9%), course attendance by 43 (27.6%), simulator availability by 46 (29.5%), and assistant participation by 76 (48.7%); 16 respondents (10.3%) reported a supervised robotic procedure or procedural step (Table 1; Fig. 1). All 56 respondents reporting training access were based in centres with RAS.

Table 1.

Study population and independent indicators of robotic training exposure

Characteristic	Overall (N = 156)
Panel A. Study population and independent indicators of RAS exposure
Residency year, median [IQR]	4 [3–6]
Senior residents, years 4–6	102/156 (65.4%)
Clinical centre	51/156 (32.7%)
Regional/voivodeship centre	56/156 (35.9%)
District/county centre	27/156 (17.3%)
Other centre	22/156 (14.1%)
Male sex	134/156 (85.9%)
RAS available in training centre	110/156 (70.5%)
Self-reported access to any robotic training	56/156 (35.9%)
Participation in any robotic course	43/156 (27.6%)
Robotic simulator available	46/156 (29.5%)
Assistant participation	76/156 (48.7%)
Supervised robotic procedure or procedural step	16/156 (10.3%)
Current self-rated skill at least satisfactory	22/156 (14.1%)
Expected self-rated skill at least satisfactory	43/156 (27.6%)
Training satisfaction at least satisfactory	36/156 (23.1%)

Outcome	Junior(n=54)	Senior(n=102)	p	Non-clinical centre(n=105)	Clinical centre(n=51)	p
Panel B. Exploratory subgroup comparisons by residency seniority and training-centre type
RAS available	37/54 (68.5%)	73/102 (71.6%)	0.691	63/105 (60.0%)	47/51 (92.2%)	<0.001
Robotic course attendance	9/54 (16.7%)	34/102 (33.3%)	0.027	28/105 (26.7%)	15/51 (29.4%)	0.719
Robotic simulator available	10/54 (18.5%)	36/102 (35.3%)	0.029	22/105 (21.0%)	24/51 (47.1%)	<0.001
Self-reported access to any robotic training	19/54 (35.2%)	37/102 (36.3%)	0.893	32/105 (30.5%)	24/51 (47.1%)	0.043
Assistant participation	26/54 (48.1%)	50/102 (49.0%)	0.917	41/105 (39.0%)	35/51 (68.6%)	0.001
Supervised robotic procedure or procedural step	1/54 (1.9%)	15/102 (14.7%)	0.011	6/105 (5.7%)	10/51 (19.6%)	0.011
Current self-rated skill at least satisfactory	3/54 (5.6%)	19/102 (18.6%)	0.029	9/105 (8.6%)	13/51 (25.5%)	0.004
Expected self-rated skill at least satisfactory	14/54 (25.9%)	29/102 (28.4%)	0.739	22/105 (21.0%)	21/51 (41.2%)	0.008
Training satisfaction at least satisfactory	16/54 (29.6%)	20/102 (19.6%)	0.158	24/105 (22.9%)	12/51 (23.5%)	0.926

Values are n/N (%) unless otherwise stated. Junior residents were in years 1–3 and senior residents in years 4–6. Non-clinical centres combine regional/voivodeship, district/county, and other categories. p values are from Pearson chi-square or Fisher exact tests, as appropriate. Variables are independent indicators rather than sequential stages; subgroup comparisons were secondary and unadjusted

IQR, interquartile range; RAS, robot-assisted surgery

Fig. 1.

Reported prevalence of independent survey indicators of institutional context, educational resources and access, and clinical participation in robot-assisted surgery. Items were not designed as sequential stages; differences in bar length are not transition or attrition rates

Variation by residency seniority and training-centre type

Senior residents more often reported robotic course attendance (33.3% vs. 16.7%, p = 0.027), simulator availability (35.3% vs. 18.5%, p = 0.029), a supervised robotic procedure or procedural step (14.7% vs. 1.9%, Fisher p = 0.011), and current self-rated skill of at least satisfactory (18.6% vs. 5.6%, Fisher p = 0.029). RAS availability, training access, assistant participation, expected skill, and satisfaction did not differ significantly by seniority (Table 1).

Compared with non-clinical centres, clinical centres more often reported RAS availability (92.2% vs. 60.0%, p < 0.001), simulator availability (47.1% vs. 21.0%, p < 0.001), training access (47.1% vs. 30.5%, p = 0.043), assistant participation (68.6% vs. 39.0%, p = 0.001), a supervised robotic procedure or procedural step (19.6% vs. 5.7%, Fisher p = 0.011), current skill of at least satisfactory (25.5% vs. 8.6%, p = 0.004), and expected skill of at least satisfactory (41.2% vs. 21.0%, p = 0.008). Course attendance and satisfaction were similar (Table 1).

Factors associated with training access and participation

Because all training-access responses occurred in centres with RAS, the exploratory access model was restricted to those 110 respondents. Robotic course attendance (aOR 5.38, 95% CI 2.05–14.14; p < 0.001) and simulator availability (aOR 3.90, 95% CI 1.54–9.89; p = 0.004) were associated with access, whereas residency year, centre type, and sex were not (Table 2; Fig. 3).

Table 2.

Multivariable models of robotic training outcomes

Outcome / model	Predictor	Adjusted OR (95% CI)	p	Method; N/events
Self-reported access to any robotic training (exploratory; RAS centres only)	Residency year	0.80 (0.59–1.07)	0.133	Logistic; 110/56
	Clinical centre	0.92 (0.38–2.19)	0.845
	Male sex	2.43 (0.69–8.50)	0.166
	Any robotic course	5.38 (2.05–14.14)	< 0.001
	Robotic simulator available	3.90 (1.54–9.89)	0.004
Assistant participation	Residency year	0.98 (0.77–1.24)	0.840	Logistic; 156/76
	Clinical centre	2.99 (1.37–6.51)	0.006
	Male sex	0.42 (0.15–1.20)	0.105
	Self-reported access to any robotic training	6.66 (3.06–14.50)	< 0.001
Supervised robotic procedure or procedural step (exploratory; training-access respondents only)	Residency year	2.11 (1.30–3.97)	0.001	Firth logistic; 56/16
	Clinical centre	3.41 (0.97–13.62)	0.056
	Male sex	1.18 (0.20–8.63)	0.858
Current self-rated robotic skill	Residency year	1.46 (1.16–1.84)	0.001	Ordinal logistic; 156
	Clinical centre	2.41 (1.22–4.79)	0.012
	Male sex	0.86 (0.33–2.21)	0.756
	Self-reported access to any robotic training	12.30 (5.88–25.73)	< 0.001
Expected end-of-training self-rated robotic skill	Residency year	1.01 (0.83–1.23)	0.890	Ordinal logistic; 156
	Clinical centre	2.21 (1.16–4.24)	0.016
	Male sex	1.60 (0.66–3.89)	0.297
	Self-reported access to any robotic training	12.15 (6.00–24.61)	< 0.001
Robotic training satisfaction at least satisfactory	Residency year	0.76 (0.57–1.02)	0.070	Logistic; 156/36
	Clinical centre	0.61 (0.23–1.58)	0.306
	Male sex	0.78 (0.22–2.80)	0.705
	Self-reported access to any robotic training	17.01 (6.35–45.57)	< 0.001

For ordinal models, the estimate is the common aOR for being in a higher response category. Covariates were selected a priori. The training-access model was restricted to respondents reporting RAS availability because no access was reported without institutional RAS. The model for a supervised robotic procedure or procedural step was restricted to respondents reporting training access because all events occurred in that group. Both conditional models are exploratory. Firth confidence intervals and p values are profile penalised-likelihood estimates. Full diagnostics are provided in Online Resource 1, Table S7

aOR, adjusted odds ratio; CI, confidence interval; RAS, robot-assisted surgery

Fig. 3.

Selected adjusted associations with robotic training access, participation, self-rated skill, and satisfaction. Training-access estimates are restricted to respondents in RAS centres (N = 110); the Firth model for a supervised robotic procedure or procedural step is restricted to respondents with training access (N = 56; 16 events). The vertical dashed line denotes an odds ratio of 1. Filled circles indicate p < 0.05 and open circles p ≥ 0.05. Ordinal-model estimates are common adjusted odds ratios

Training access was associated with assistant participation (aOR 6.66, 95% CI 3.06–14.50; p < 0.001), and clinical-centre training was independently associated with assistant participation (aOR 2.99, 95% CI 1.37–6.51; p = 0.006). All 16 reports of a supervised robotic procedure or procedural step occurred among the 56 respondents with training access. In the exploratory conditional Firth model, each additional residency year was associated with higher odds of this outcome (aOR 2.11, 95% CI 1.30–3.97; profile-likelihood p = 0.001); the clinical-centre estimate was imprecise (aOR 3.41, 95% CI 0.97–13.62; p = 0.056). Respondent-level cross-tabulations showed assistant participation in 76.8% with versus 33.0% without training access and the supervised outcome in 28.6% versus 0%, respectively (Online Resource 1, Table S6).

Self-rated skill and satisfaction

Training access was associated with a higher current self-rated robotic skill category (common aOR 12.30, 95% CI 5.88–25.73; p < 0.001) and higher expected end-of-training skill (common aOR 12.15, 95% CI 6.00–24.61; p < 0.001). Later residency year (common aOR 1.46, 95% CI 1.16–1.84; p = 0.001) and clinical-centre training (common aOR 2.41, 95% CI 1.22–4.79; p = 0.012) were associated with current skill; clinical-centre training was also associated with expected skill (common aOR 2.21, 95% CI 1.16–4.24; p = 0.016) (Table 2; Fig. 3).

Current skill was rated as none by 83 (53.2%), very poor by 20 (12.8%), poor by 31 (19.9%), satisfactory by 16 (10.3%), good by 6 (3.8%), and very good by none. Thus, 22 (14.1%) rated current skill as at least satisfactory. Expected end-of-training skill was at least satisfactory in 43 (27.6%). Median skill increased from 0 [IQR 0–2] to 1 [IQR 0–3] (Wilcoxon signed-rank p < 0.001); 52 respondents selected a higher expected category, 103 the same category, and one a lower category (Fig. 2; Online Resource 1, Table S1).

Fig. 2.

Distribution of current and expected self-rated robotic skill and satisfaction with robotic training. Panel a presents six ordered skill categories (none, very poor, poor, satisfactory, good, and very good); panel b presents five satisfaction categories. Ratings were self-reported and were not objective proficiency measures

Robotic training satisfaction was very low: 50 (32.1%) were very dissatisfied, 70 (44.9%) dissatisfied, 25 (16.0%) selected the neutral/satisfactory category, and 11 (7.1%) were satisfied; none were very satisfied. Overall, 120 (76.9%) were dissatisfied or very dissatisfied, while 36 (23.1%) rated training as at least satisfactory. Training access was strongly associated with at least satisfactory satisfaction (aOR 17.01, 95% CI 6.35–45.57; p < 0.001) (Table 2; Fig. 2). Model diagnostics showed no material multicollinearity (maximum VIF 1.13 across all models). Global likelihood-ratio tests were significant for each conventional logistic and ordinal model (all p < 0.001). AUC values were 0.767 for training access, 0.791 for assistant participation, 0.805 for supervised experience, and 0.840 for satisfaction; corresponding Brier scores were 0.197, 0.188, 0.157, and 0.128. The proportional-odds assumption was not rejected for current skill (p = 0.738) or expected skill (p = 0.254) (Online Resource 1, Table S7).

Procedural preferences, comparative advantages, and perceived barriers

Assuming equal access, residents preferred RAS for radical prostatectomy (93.6%), radical cystectomy (84.0%), partial nephrectomy (71.2%), and pyeloplasty (62.2%), whereas laparoscopy was preferred for radical nephrectomy (70.5%) (Online Resource 1, Table S2). RAS was perceived to have an advantage in ease of performing the procedure (82.1%), ease of learning (78.2%), and operating time (53.8%), whereas laparoscopy was favoured for procedure cost (85.9%), access to training (75.0%), training cost (70.5%), and access to tutors (64.1%). Limited availability (42.3%), lack of time (22.4%), and lack of a mentor (17.9%) were the most frequently selected primary barriers. Workload or working hours (77.4%), mentoring (72.3%), and workplace atmosphere (70.3%) were most frequently rated as high-impact broader negative training factors (Online Resource 1, Table S3; Fig. 4).

Fig. 4.

Perceived comparative advantages, barriers to robotic training, and broader negative training factors. Panel a presents comparative ratings: the left side indicates an advantage of laparoscopy, the centre no clear advantage, and the right side an advantage of RAS. Panel b shows the single primary barrier selected by each respondent. Panel c shows the proportion assigning high impact (scores 4–5) to broader factors, with exact percentages and median [IQR] ratings displayed; n = 155 for panel c

Discussion

The principal finding was the differing prevalence and respondent-level overlap of independent indicators of robotic infrastructure, educational resources, and clinical participation. RAS availability was reported by 70.5%, training access by 35.9%, and assistant participation by 48.7%, while 10.3% reported a supervised robotic procedure or procedural step. These values do not describe sequential transitions. Nevertheless, the cross-tabulations are informative: all respondents with training access were in RAS centres, but only 50.9% of respondents in RAS centres reported access; among those with access, 76.8% had assisted and 28.6% reported the supervised outcome. The study therefore quantifies educational integration rather than a training pathway.

This narrower contribution is relevant in Poland’s high-adoption setting. In 2025, 93 hospitals performed 24,585 robot-assisted operations, including 15,032 urological procedures and 13,231 radical prostatectomies [9]. The current national programme includes a 2-week robotic-urology rotation covering system safety, patient qualification, docking, assisting, simulator work when available, and video review, but it does not specify a minimum number of console cases or modular steps [10]. Some respondents entered residency under earlier programme versions, and our survey did not audit curriculum compliance. The results should therefore be read as a national snapshot during rapid implementation, not as proof that every centre failed to deliver a prescribed pathway.

Course attendance and simulator availability were independently associated with reported training access within RAS centres. Training access, in turn, was associated with assistant participation, higher self-rated skill, and higher satisfaction, while clinical-centre training was associated with assistant participation and skill. These relationships are educationally plausible and consistent with curricula combining simulation and supervised clinical exposure [1, 2, 14–18]. However, their cross-sectional nature, conceptual overlap, and common method of self-report preclude causal interpretation. The large access coefficients may partly reflect related components of the same educational environment.

The supervised outcome warrants particular restraint. Only 16 respondents reported a supervised robotic procedure or procedural step, and all were within the training-access group. The conditional Firth analysis suggested an association with later residency year, while the clinical-centre estimate was large but imprecise. These findings are hypothesis-generating. Progression to supervised operating may appropriately depend on interest, prerequisite proficiency, seniority, case mix, programme capacity, and faculty assessment. Because applications, eligibility, selection, available places, and reasons for individual decisions were not collected, the low prevalence cannot be interpreted as evidence of inappropriate exclusion.

Self-rated robotic skill was low: only 14.1% rated current skill as at least satisfactory and 27.6% expected to reach that level by the end of residency. Training access showed strong associations with both outcomes, and later year and clinical-centre training were associated with current skill. These are subjective perceptions rather than objective proficiency. Self-selection into training, greater confidence among exposed residents, and shared-response bias may all contribute; future studies should use simulator metrics, structured operative assessment, console time, and repeated modular performance [13, 17, 19, 20, 21]. This is consistent with recent expert consensus emphasizing structured, competency-based robotic curricula with objective assessment and progressive skills acquisition rather than reliance on self-reported confidence alone [22].

Low satisfaction remained prominent: 76.9% were dissatisfied or very dissatisfied. Training access showed a strong association with at least satisfactory satisfaction in the adjusted model. This does not prove that access causes satisfaction; residents with favourable experiences may be more likely to report access, and both variables were measured in the same survey. The instrument’s broad content makes participation driven solely by pre-existing interest in robotics less likely than in a robotics-only questionnaire, but it cannot eliminate non-response bias or differential participation by residents with especially positive or negative experiences.

Procedural preferences provide context rather than evidence of clinical superiority. Respondents preferred RAS for radical prostatectomy, partial nephrectomy, pyeloplasty, and cystectomy, while laparoscopy remained preferred for radical nephrectomy. RAS was perceived as easier to perform and learn, whereas laparoscopy was favoured for cost, training access, and tutor availability. These perceptions support maintaining laparoscopic education alongside robotic curricula, especially where technology, case volume, or faculty capacity varies.

Several limitations remain. The cross-sectional, self-reported design cannot establish temporal order, causality, curriculum effectiveness, or objective competence. The 40.0% completion proportion creates non-response bias; no characteristics of non-respondents were available, so its direction and magnitude are unknown. Although the questionnaire addressed laparoscopy and residency quality more broadly, this reduces rather than excludes selection related to robotics or surgical education. Thirty-nine initiated questionnaires were incomplete and excluded. Centre identifiers and geography were not collected, preventing clustering analyses and comparison of response proportions across centre categories. The broad access item may overlap with courses and simulator availability, and all key outcomes share a common self-report method. Reports of a supervised robotic procedure or procedural step were rare; the item did not identify whether the experience involved a complete operation or an individual step, and the outcome was analysed exploratorily. Trainee interest, applications, eligibility, selection, programme capacity, case numbers, console time, repeated exposure, and objective proficiency were not measured.

Conclusions

In this national respondent sample, institutional RAS availability was reported more often than robotic training access, while only 10.3% reported a supervised robotic procedure or procedural step. These independent indicators should not be interpreted as sequential stages. Course attendance and simulator availability were associated with access within RAS centres, while access was associated with assistant participation, self-rated skill, and satisfaction. The supervised outcome remained uncommon, and its conditional analysis was exploratory. Prospective national audits should measure trainee interest and selection, programme capacity, console time, modular participation, repeated supervised performance, and objective proficiency, while the present findings are interpreted in light of non-response and self-report limitations.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

B.B. conceived and designed the study, developed the questionnaire, coordinated data collection, performed the statistical analysis, and wrote the main manuscript text. M.J. contributed to study design, data collection, and manuscript preparation. R.B. contributed to data collection and interpretation of the results. B.R. contributed to data collection and manuscript revision. Ł.M. contributed to data analysis and interpretation. R.B.D. contributed to the methodological and statistical review. M.W.J. contributed to data collection and manuscript revision. M.W.P. contributed to data collection and critical revision of the manuscript. M.T. contributed to study supervision and critical revision of the manuscript. P.J. supervised the study and critically revised the manuscript. All authors reviewed and approved the final manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The datasets analysed during the current study are not publicly available due to the potential risk of participant identification in this relatively small professional population but are available from the corresponding author on reasonable request.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethics approval

Under applicable Polish regulations and institutional policy, formal ethics committee review was not required for this anonymous educational survey.

Consent to participate

Participation was voluntary, and submission of the anonymous questionnaire indicated informed consent to participate.

Consent for publication

Not applicable. Only aggregate, non-identifiable data are reported.