Socioeconomic differences in prostate cancer diagnosis and stage following the introduction of Rapid Access Prostate Clinics in Ireland Authors: C. Haugh, S. Walter, P. Sweeney, J. McDevitt, D. Murray, M.T. Redaniel Published: 2026 ABSTRACT BackgroundRapid Access Prostate Clinics (RAPCs) were introduced in Ireland in 2009 to centralise and standardise prostate cancer (PCa) diagnosis within designated cancer centres. While RAPCs aim to improve diagnostic organisation following general practitioner referral, access remains influenced by socioeconomic differences in prostate-specific antigen testing and referral patterns. We examined changes in PCa incidence and stage at diagnosis following RAPC implementation and whether trends differed across socioeconomic groups.Patients and methodsUsing population-based data from the National Cancer Registry Ireland, we identified men aged 50-70 years diagnosed with PCa between 2000 and 2022. Socioeconomic status was assigned using the Pobal Haase–Pratschke Deprivation Index. Interrupted time-series analyses using negative binomial regression compared trends before (2000-2008) and after (2009-2022) RAPC implementation. Analyses were conducted by deprivation quintile and tumour stage, with sensitivity analyses restricted to cancer centres.ResultsA total of 44 127 cases were identified. Before RAPC implementation, quarterly incidence estimates increased across all socioeconomic groups. Following implementation, incidence trends stabilised, with a decline in the most deprived group. Early-stage (I and II) incidence was most affected, stabilising or decreasing, particularly among more deprived groups, while stages III and IV incidence rates showed modest increases following implementation, with limited socioeconomic variation.ConclusionsRAPC implementation coincided with stabilisation of PCa incidence following a period of rising diagnoses. However, it did not reduce late-stage presentations or eliminate socioeconomic differences in stage at presentation. These findings suggest that centralisation of diagnostic services alone is insufficient in achieving equity in PCa detection. Introduction Urological cancers represent a major global health burden, with prostate cancer (PCa) accounting for an estimated 1.4 million new diagnoses in 2022.1 In Ireland, PCa comprises ∼30% of all invasive cancers in males between 2020 and 2022. The clinical and economic impact is considerable: PCa alone generated an estimated €45.6 million in health care expenditure in 2010, reflecting costs spanning diagnosis, treatment, surveillance, and end-of-life care.2 With rising incidence and health care expenditure in Ireland, the current economic burden is likely substantially higher.3 Morbidity related to diagnosis and treatment, including urinary incontinence, sexual dysfunction, and fatigue,4 further underscores the importance of timely and effective diagnostic pathways. Socioeconomic deprivation shapes access to PCa diagnostics in Ireland.5 Individuals with suspected cancer may be assessed through private hospitals, public hospitals, or designated cancer centres—public hospitals that have been upgraded to deliver comprehensive, multidisciplinary cancer care in line with national clinical guidance.6 Men living in more deprived groups are more likely to rely on public hospital pathways, while less deprived men more often use private services, reflecting differences in insurance coverage and health care access.7 These differences in where men enter the diagnostic pathway can influence stage at diagnosis and must be considered when assessing systemwide reforms. The National Cancer Strategy 2006 prioritised the centralisation of cancer care with a particular focus on centralisation of confirmatory diagnosis, treatment planning and surgery to improve outcomes,8 leading to the establishment of the National Cancer Control Programme (NCCP) in 2007 and designation of eight adult cancer centres by 2009.9 This reconfiguration of services coincided with the introduction of national general practitioner (GP) referral guidelines for suspected PCa, aimed at standardising referral pathways to specialist care.10 As part of this centralised model, the NCCP introduced Rapid Access Prostate Clinics (RAPCs) within cancer centres in 2009 to provide timely, consultant-led diagnostic pathways for men with PCa, particularly those aged 50-70 years or younger if at elevated risk.10 Referral to RAPCs was guided by NCCP recommendations for men with suspected PCa, including those with elevated prostate-specific antigen (PSA) levels, abnormal digital rectal examination findings, or other clinical features considered suspicious for malignancy. Implementation of RAPCs was phased nationwide between 2009 and 2011. RAPCs were designed to replace more fragmented diagnostic processes in public hospitals by streamlining referrals and standardising clinical assessment, with the aim of improving diagnostic timeliness, as observed in evaluations of comparable rapid access clinics.11 Early RAPC evaluations demonstrated improved pathological characteristics of men presenting with PCa.12 A recent National Cancer Registry Ireland (NCRI) report13 documented a clear centralisation of PCa diagnosis towards designated cancer centres after RAPC implementation, accompanied by a stabilisation of national incidence rates but a persistent increase in stage III and IV presentations. Building on the NCRI report, this study examines trends in PCa incidence and stage at presentation by socioeconomic deprivation quintile, both nationally and within cancer centres. We aim to determine whether there are disparities on the impact of centralised diagnostic reform on diagnostic patterns across socioeconomic strata. Materials and methods Data source and study population We used data from the NCRI, the state agency responsible for collecting and analysing data from all patients with cancer in the Republic of Ireland. The registry dataset is known to be of high quality as documented on annual submissions to the European Cancer Information System.14 From the NCRI dataset, we identified men aged 50-70 years who were diagnosed with PCa (indicated by International Classification of Diseases-10 code C61 in their records) between 2000 and 2022. Analyses were restricted by age to reflect the target population for the RAPC referral pathway. Tumour stage has been defined using the TNM (tumour–node–metastasis) classification system.15 Since 2014, NCRI uses TNM7, with TNM4 and TNM5 used in previous years. For consistency, cases diagnosed between 2000 and 2013 were recoded to follow TNM7 conventions using available clinical T, N, and M information.15 TNM7 stage groups were assigned according to standard TNM7 criteria, with T1/T2aN0M0 classified as stage I, T2b/T2cN0M0 as stage II, T3N0M0 as stage III, and T4 disease, N1 disease, or M1 disease classified as stage IV. Socioeconomic status (SES) was assigned using the Pobal Haase–Pratschke (HP) Deprivation Index linked to address at diagnosis and categorised into quintiles (Q1 least deprived to Q5 most deprived).16 The HP Index is an area-based measure of socioeconomic deprivation derived from Irish census data that incorporates indicators of socioeconomic disadvantage from three categories (demographic, social class, and labour market situation). Deprivation scores were assigned according to the Electoral Division, corresponding to the patient’s address at diagnosis. Electoral Divisions had a median population of 675 people (interquartile range 372-1525) in Census 2022. Statistical analysis Quarterly counts of newly diagnosed PCa cases (incident cases) were estimated, with 2000-2008 defined as the preimplementation period and 2009-2022 as the implementation period. Annual population estimates for men aged 50-70 years in Ireland were used to derive quarterly person-year denominators, with annual population counts divided equally across the four quarters of each calendar year. Analyses were not age-standardised within the 50-70–year age range, as the age distribution of cases was similar before and after RAPC implementation, with differences of less than one percentage point across all age categories. An interrupted time-series analysis was conducted using negative binomial regression applied to quarterly case numbers.17 Negative binomial regression was used because the corresponding Poisson model exhibited significant overdispersion (P < 0.001). Residual autocorrelation was assessed using correlograms of model residuals and no meaningful autocorrelation was identified. All models were fitted using robust standard errors. Time was coded as a continuous quarterly variable centred on the RAPC intervention point, with Q1 2009 assigned a value of 0 and values increasing or decreasing by one for each subsequent or preceding quarter, respectively. A period indicator was coded 0 for the preimplementation period and 1 for the implementation period. The base model included period, time, and a period-by-time interaction term, allowing estimation of the immediate level change associated with RAPC implementation and the change in incidence trend (slope change) following implementation. The logarithm of person-years was included as an offset. Deprivation-specific population denominators were not available for the study period and therefore were not incorporated into the analyses. Results are reported as incidence rate ratios (IRRs) with 95% confidence intervals (CIs). To examine the possible impact of RAPC introduction on incidence trends in socioeconomic groups, a three-way interaction between time, period, and deprivation quintile was included in the base model, allowing estimation of deprivation group-specific slope changes. To examine the possible impact of RAPC on stage at diagnosis, a three-way interaction between time, period, and stage was included in the base model, allowing estimation of stage-specific slope changes. Deprivation-stratified interrupted time series models were fitted to characterise stage-specific patterns within each quintile. Wald tests were used to evaluate the overall heterogeneity of RAPC-related changes in incidence trends across deprivation quintiles within each stage, based on the deprivation-by-time interaction terms. As state-supported RAPCs operate within designated cancer centres, these analyses were also conducted on cases diagnosed in cancer centres only. Trends within cancer centres reflect both underlying incidence patterns and the progressive centralisation of PCa diagnostic activity into designated cancer centres following RAPC implementation. Sensitivity analyses Several sensitivity analyses were undertaken to assess the robustness of the findings. Firstly, to account for the phased implementation of RAPCs between 2009 and 2011, all analyses were repeated using alternative intervention dates of Q1 2010 and Q1 2011. A further analysis excluded the implementation period (2009-2011) and defined Q1 2012 as the intervention point. To assess the potential impact of disruption to cancer diagnostic services during the coronavirus disease (COVID-19) pandemic, analyses were repeated after excluding observations from 2020 to 2022. To evaluate the potential influence of changes in staging completeness over time, stage-specific analyses were repeated with unstaged tumours included as a separate category. Lastly, potential seasonal variation was evaluated by including calendar quarter as a categorical variable in sensitivity analyses. Results from all sensitivity analyses are presented in Supplementary Tables S4-S15 (available at https://doi.org/10.1016/j.esmorw.2026.100781). Results Descriptive analysis A total of 44 127 PCa cases were diagnosed among individuals aged 50-70 years between 2000 and 2022. Of these, 11 564 were diagnosed in the preimplementation period (2000-2008) and 32 563 in the implementation period (2009-2022) (Table 1). Quarterly case numbers rose following implementation, corresponding to an increase in crude quarterly incidence rates from 324 to 450 per 100 000 person-years.Table 1Sociodemographic characteristics of patients with PCa aged 50-70 diagnosed between 2000 and 2022VariableOverall (2000-2022)Preimplementation (2000-2008)Postimplementation (2009-2022)N44 12711 56432 563Deprivation quintile17924 (18%)2069 (17.9%)5855 (18%)27557 (17.1%)2010 (17.4%)5547 (17%)38000 (18.1%)2008 (17.4%)5992 (18.4%)48481 (19.2%)2271 (19.6%)6210 (19.1%)58530 (19.3%)2223 (19.2%)6307 (19.4%)Unknown3635 (8.2%)983 (8.5%)2652 (8.1%)Stage at diagnosisI19 308 (43.8%)4321 (37.4%)14 987 (46%)II11 675 (26.5%)3197 (27.6%)8478 (26%)III6862 (15.6%)1275 (11%)5587 (17.2%)IV3140 (7.1%)851 (7.4%)2289 (7%)Unknown3142 (7.1%)1920 (16.6%)1222 (3.8%)Hospital type of diagnosisCancer centre19 257 (43.6%)3332 (28.8%)15 925 (48.9%)Public hospital10 451 (23.7%)4925 (42.6%)5526 (17%)Private hospital14 419 (32.7%)3307 (28.6%)11 112 (34.1%)Quarterly case numbersMean (range)480 (179-691)321 (179-437)581 (384-691)Quarterly crude incidence rates per 100 000 person-yearsMean (range)400.9 (203.5-547.4)324 (203.5-415.7)450.3 (270.7-547.4)PCa, prostate cancer. Table 1Sociodemographic characteristics of patients with PCa aged 50-70 diagnosed between 2000 and 2022VariableOverall (2000-2022)Preimplementation (2000-2008)Postimplementation (2009-2022)N44 12711 56432 563Deprivation quintile17924 (18%)2069 (17.9%)5855 (18%)27557 (17.1%)2010 (17.4%)5547 (17%)38000 (18.1%)2008 (17.4%)5992 (18.4%)48481 (19.2%)2271 (19.6%)6210 (19.1%)58530 (19.3%)2223 (19.2%)6307 (19.4%)Unknown3635 (8.2%)983 (8.5%)2652 (8.1%)Stage at diagnosisI19 308 (43.8%)4321 (37.4%)14 987 (46%)II11 675 (26.5%)3197 (27.6%)8478 (26%)III6862 (15.6%)1275 (11%)5587 (17.2%)IV3140 (7.1%)851 (7.4%)2289 (7%)Unknown3142 (7.1%)1920 (16.6%)1222 (3.8%)Hospital type of diagnosisCancer centre19 257 (43.6%)3332 (28.8%)15 925 (48.9%)Public hospital10 451 (23.7%)4925 (42.6%)5526 (17%)Private hospital14 419 (32.7%)3307 (28.6%)11 112 (34.1%)Quarterly case numbersMean (range)480 (179-691)321 (179-437)581 (384-691)Quarterly crude incidence rates per 100 000 person-yearsMean (range)400.9 (203.5-547.4)324 (203.5-415.7)450.3 (270.7-547.4)PCa, prostate cancer. Sociodemographic characteristics of patients with PCa aged 50-70 diagnosed between 2000 and 2022 PCa, prostate cancer. The distribution of cases across deprivation quintiles remained broadly consistent over time, with each quintile comprising ∼17%-19% of cases and the most deprived quintile (Q5) accounting for 19.3% overall. Stage distribution shifted markedly across periods. The proportion of stage I diagnoses increased from 37.4% before implementation to 46% thereafter. Stage II cancers comprised approximately one-quarter of cases in both periods, while stage III cancers increased from 11% preimplementation to 17.2% in the implementation period. Stage IV cancers remained stable at around 7% and unknown stage declined sharply from 16.6% to 3.8%. Impact of RAPC introduction on incidence trends in socioeconomic quintiles There were clear differences in incidence trends across socioeconomic deprivation quintiles both nationally and within cancer centres (Figure 1 and Table 2). Across all hospitals, PCa incidence increased in all deprivation quintiles during the preimplementation period, with quarterly increases ranging from 1.04% in the least deprived quintile to 1.96% in the second least deprived group and 1.67% in the most deprived quintile.Figure 1Trends in incidence pre- and post-RAPC implementation by deprivation quintile, nationally, and in cancer centres. RAPC, Rapid Access Prostate Clinic.Table 2Relative trend in incidence rates (slope IRR) pre- and post-implementation by deprivation quintile, nationally and in cancer centres onlyHospital typeDeprivation quintilePercentage change in incidence rate per quarterSlope IRRChange in slopeImmediate level changePreimplementation periodImplementation periodPreimplementation periodImplementation period% (95% CI)% (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)All hospitals11.04 (0.52-1.57)0.18 (−0.04 to 0.40)1.0104 (1.0052-1.0157)1.0018 (0.9996-1.0040)0.9914 (0.9859-0.9971)1.0928 (0.9702-1.2309)21.96 (1.48-2.44)−0.03 (−0.23 to 0.18)1.0196 (1.0148-1.0244)0.9997 (0.9977-1.0018)0.9806 (0.9755-0.9856)0.9781 (0.8797-1.0875)31.49 (1.08-1.91)−0.02 (−0.23 to 0.19)1.0149 (1.0108-1.0191)0.9998 (0.9977-1.0019)0.9851 (0.9805-0.9897)1.1361 (1.0342-1.2480)41.61 (1.22-1.99)−0.20 (−0.42 to 0.01)1.0161 (1.0122-1.0199)0.9980 (0.9958-1.0001)0.9822 (0.9779-0.9865)1.0774 (0.9731-1.1929)51.67 (1.19-2.15)−0.45 (−0.62 to −0.28)1.0167 (1.0119-1.0215)0.9955 (0.9938-0.9972)0.9822 (0.9779-0.9865)1.0774 (0.9731-1.1929)Cancer centres12.20 (1.06-3.36)0.37 (0.01-0.72)1.0220 (1.0106-1.0336)1.0037 (1.0001-1.0072)0.9820 (0.9705-0.9936)1.4033 (1.0873-1.8111)22.53 (1.79-3.28)0.69 (0.36-1.01)1.0253 (1.0179-1.0328)1.0069 (1.0036-1.0101)0.9820 (0.9742-0.9898)1.2252 (1.0504-1.4290)32.40 (1.50-3.31)0.62 (0.32-0.93)1.0240 (1.0150-1.0331)1.0062 (1.0032-1.0093)0.9827 (0.9735-0.9919)1.3417 (1.1125-1.6182)42.54 (1.58-3.52)0.41 (0.03-0.80)1.0254 (1.0158-1.0352)1.0041 (1.0003-1.0080)0.9792 (0.9693-0.9893)1.4100 (1.1350-1.7515)54.37 (3.55-5.19)−0.13 (−0.38 to 0.13)1.0437 (1.0355-1.0519)0.9987 (0.9962-1.0013)0.9792 (0.9693-0.9893)1.4100 (1.1350-1.7515)CI, confidence interval; IRR, incidence rate ratio. Trends in incidence pre- and post-RAPC implementation by deprivation quintile, nationally, and in cancer centres. RAPC, Rapid Access Prostate Clinic. Table 2Relative trend in incidence rates (slope IRR) pre- and post-implementation by deprivation quintile, nationally and in cancer centres onlyHospital typeDeprivation quintilePercentage change in incidence rate per quarterSlope IRRChange in slopeImmediate level changePreimplementation periodImplementation periodPreimplementation periodImplementation period% (95% CI)% (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)All hospitals11.04 (0.52-1.57)0.18 (−0.04 to 0.40)1.0104 (1.0052-1.0157)1.0018 (0.9996-1.0040)0.9914 (0.9859-0.9971)1.0928 (0.9702-1.2309)21.96 (1.48-2.44)−0.03 (−0.23 to 0.18)1.0196 (1.0148-1.0244)0.9997 (0.9977-1.0018)0.9806 (0.9755-0.9856)0.9781 (0.8797-1.0875)31.49 (1.08-1.91)−0.02 (−0.23 to 0.19)1.0149 (1.0108-1.0191)0.9998 (0.9977-1.0019)0.9851 (0.9805-0.9897)1.1361 (1.0342-1.2480)41.61 (1.22-1.99)−0.20 (−0.42 to 0.01)1.0161 (1.0122-1.0199)0.9980 (0.9958-1.0001)0.9822 (0.9779-0.9865)1.0774 (0.9731-1.1929)51.67 (1.19-2.15)−0.45 (−0.62 to −0.28)1.0167 (1.0119-1.0215)0.9955 (0.9938-0.9972)0.9822 (0.9779-0.9865)1.0774 (0.9731-1.1929)Cancer centres12.20 (1.06-3.36)0.37 (0.01-0.72)1.0220 (1.0106-1.0336)1.0037 (1.0001-1.0072)0.9820 (0.9705-0.9936)1.4033 (1.0873-1.8111)22.53 (1.79-3.28)0.69 (0.36-1.01)1.0253 (1.0179-1.0328)1.0069 (1.0036-1.0101)0.9820 (0.9742-0.9898)1.2252 (1.0504-1.4290)32.40 (1.50-3.31)0.62 (0.32-0.93)1.0240 (1.0150-1.0331)1.0062 (1.0032-1.0093)0.9827 (0.9735-0.9919)1.3417 (1.1125-1.6182)42.54 (1.58-3.52)0.41 (0.03-0.80)1.0254 (1.0158-1.0352)1.0041 (1.0003-1.0080)0.9792 (0.9693-0.9893)1.4100 (1.1350-1.7515)54.37 (3.55-5.19)−0.13 (−0.38 to 0.13)1.0437 (1.0355-1.0519)0.9987 (0.9962-1.0013)0.9792 (0.9693-0.9893)1.4100 (1.1350-1.7515)CI, confidence interval; IRR, incidence rate ratio. Relative trend in incidence rates (slope IRR) pre- and post-implementation by deprivation quintile, nationally and in cancer centres only CI, confidence interval; IRR, incidence rate ratio. Following RAPC implementation, incidence trends stabilised across deprivation quintiles. In quintiles 1-4, postimplementation changes were small in magnitude, with CIs including the null, indicating a flattening of previously increasing incidence trends. A modest decline was observed in the most deprived quintile (−0.45% per quarter, 95% CI −0.62% to −0.28%), indicating a clear downwards trend relative to the preimplementation period. When analyses were restricted to cancer centres, preimplementation increases were steeper across all deprivation groups, particularly in the most deprived quintile, where incidence rose by 4.37% per quarter (95% CI 3.55% to 5.19%). During the implementation period, trends within cancer centres were markedly attenuated and largely converged across deprivation quintiles. Postimplementation changes were generally small and showed little variation across deprivation groups. Estimated quarterly changes ranged from −0.13% to 0.69%, with a slight, nonsignificant downwards trend observed in the most deprived quintile (−0.13% per quarter, 95% CI −0.38% to 0.13%). Diagnostic setting also varied by deprivation quintile, with less deprived groups more frequently diagnosed in private hospitals and more deprived groups showing steeper preimplementation increases in cancer-centre diagnoses (Supplementary Figure S1 and Table S1, available at https://doi.org/10.1016/j.esmorw.2026.100781). Impact of RAPC introduction on stage at presentation Stage-specific analyses demonstrated marked differences in incidence trends before and after RAPC implementation, both nationally and within cancer centres (Figure 2 and Supplementary Table S2, available at https://doi.org/10.1016/j.esmorw.2026.100781). Across all hospitals, stage I and stage II PCa incidence increased substantially during the preimplementation period, rising by 3.95% and 2.51% per quarter, respectively.Figure 2Trends in incidence pre- and post-RAPC implementation by TNM stage at diagnosis, nationally, and in cancer centres only. RAPC, Rapid Access Prostate Clinic; TNM, tumour–node–metastasis. Trends in incidence pre- and post-RAPC implementation by TNM stage at diagnosis, nationally, and in cancer centres only. RAPC, Rapid Access Prostate Clinic; TNM, tumour–node–metastasis. Following implementation, the rate of increase was reduced across all stages, indicating a slowing of previously rising incidence. Stage I incidence stabilised (−0.01% per quarter, 95% CI −0.23% to 0.21%), while stage II incidence declined (−0.82% per quarter, 95% CI −1.06% to −0.59%), representing the clearest reduction following implementation. In contrast, stage III incidence continued to increase during implementation, albeit at a slower rate than preimplementation, while stage IV incidence shifted from a stable preimplementation pattern to a modest positive trend. When analyses were restricted to cancer centres, preimplementation increases were consistently steeper across all stages. Stage I and stage II incidence increased by 4.89% and 5.33% per quarter, respectively, before implementation. During the implementation period, growth in incidence was reduced across all stages, with postimplementation changes small in magnitude. Stage I incidence showed a slight increase (0.36% per quarter), while stage II incidence flattened. Both stage III and stage IV cancers continued to rise during implementation within cancer centres, with quarterly increases of 0.84% and 0.82%, respectively. Across all stages, slope IRRs were close to 1, suggesting little evidence of sustained changes in incidence within cancer centres following implementation. Impact of RAPC introduction on stage at presentation—stratified by deprivation quintile Stage-specific incidence trends stratified by socioeconomic deprivation differed between the preimplementation and implementation periods, with notable divergence in postimplementation stage I trends across deprivation quintiles (Table 3 and Figure 3). Across all hospitals, stage I incidence increased in all deprivation groups before RAPC introduction, with quarterly increases of ∼3%-4%. Following implementation, stage I incidence showed no ongoing increase in deprivation quintiles 1-4, while a small decline was observed in the most deprived quintile (−0.23% per quarter, 95% CI −0.50% to 0.04%), although the CI includes the null and suggests uncertainty in this estimate.Table 3Relative trend in incidence rates (slope IRR) pre- and post-implementation by TNM stage at diagnosis, stratified by deprivation quintile subgroupsDeprivation quintileTNM 7 stagePercentage change in incidence rate per quarterSlope IRRChange in slopeImmediate level changePreimplementation periodImplementation periodPreimplementation periodImplementation period% (95% CI)% (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)1I3.03 (2.24-3.83)0.18 (−0.11 to 0.46)1.0303 (1.0224-1.0383)1.0018 (0.9989-1.0046)0.9723 (0.9643-0.9802)0.9911 (0.8342-1.1773)II1.50 (0.52-2.48)−0.56 (−0.90 to −0.21)1.0150 (1.0052-1.0248)0.9944 (0.9910-0.9979)0.9798 (0.9698-0.9899)1.2711 (1.0276-1.5723)III1.02 (−0.25 to 2.30)1.00 (0.61-1.39)1.0102 (0.9975-1.0230)1.0100 (1.0061-1.0139)0.9998 (0.9867-1.0131)1.1899 (0.9304-1.5219)IV0.11 (−1.62 to 1.87)0.75 (0.19-1.32)1.0011 (0.9838-1.0187)1.0075 (1.0019-1.0132)1.0064 (0.9882-1.0250)1.2968 (0.8664-1.9409)2I3.82 (3.17-4.48)0.11 (−0.16 to 0.39)1.0382 (1.0317-1.0448)1.0011 (0.9984-1.0039)0.9643 (0.9576-0.9710)0.8904 (0.7769-1.0204)II3.11 (2.10-4.12)−0.67 (−1.05 to −0.28)1.0311 (1.0210-1.0412)0.9933 (0.9895-0.9972)0.9634 (0.9533-0.9736)0.8682 (0.7132-1.0568)III1.74 (0.62-2.87)0.41 (0.02-0.81)1.0174 (1.0062-1.0287)1.0041 (1.0002-1.0081)0.9870 (0.9755-0.9986)1.3484 (1.0208-1.7811)IV−0.43 (−1.63 to 0.79)0.29 (−0.27 to 0.86)0.9957 (0.9837-1.0079)1.0029 (0.9973-1.0086)1.0072 (0.9938-1.0208)1.3627 (0.9516-1.9512)3I3.82 (3.14-4.50)0.11 (−0.19 to 0.40)1.0382 (1.0314-1.0450)1.0011 (0.9981-1.0040)0.9643 (0.9573-0.9713)0.9376 (0.8099-1.0854)II1.87 (1.04-2.71)−0.78 (−1.14 to −0.42)1.0187 (1.0104-1.0271)0.9922 (0.9886-0.9958)0.9740 (0.9653-0.9827)1.2854 (1.0697-1.5446)III1.42 (0.68-2.17)0.52 (0.12-0.93)1.0142 (1.0068-1.0217)1.0052 (1.0012-1.0093)0.9911 (0.9829-0.9995)1.5495 (1.2297-1.9525)IV−1.10 (−2.55 to 0.36)0.57 (−0.06 to 1.20)0.9890 (0.9745-1.0036)1.0057 (0.9994-1.0120)1.0169 (1.0008-1.0333)1.3441 (0.9117-1.9814)4I3.93 (2.99-4.88)−0.03 (−0.33 to 0.28)1.0393 (1.0299-1.0488)0.9997 (0.9967-1.0028)0.9619 (0.9527-0.9712)0.9189 (0.7735-1.0916)II2.28 (1.46-3.11)−0.85 (−1.18 to −0.53)1.0228 (1.0146-1.0311)0.9915 (0.9882-0.9947)0.9693 (0.9609-0.9779)1.0362 (0.8603-1.2480)III2.89 (1.86-3.93)0.28 (−0.15 to 0.72)1.0289 (1.0186-1.0393)1.0028 (0.9985-1.0072)0.9746 (0.9640-0.9854)1.3030 (1.0108-1.6797)IV−0.61 (−2.00 to 0.80)0.35 (−0.29 to 1.00)0.9939 (0.9800-1.0080)1.0035 (0.9971-1.0100)1.0097 (0.9942-1.0255)1.2191 (0.8459-1.7569)5I3.89 (3.04-4.75)−0.23 (−0.50 to 0.04)1.0389 (1.0304-1.0475)0.9977 (0.9950-1.0004)0.9604 (0.9520-0.9688)0.9890 (0.8403-1.1641)II2.83 (1.79-3.89)−1.16 (−1.52 to −0.80)1.0283 (1.0179-1.0389)0.9884 (0.9848-0.9920)0.9612 (0.9508-0.9717)1.1728 (0.9560-1.4386)III2.00 (0.77-3.25)0.10 (−0.26 to 0.46)1.0200 (1.0077-1.0325)1.0010 (0.9974-1.0046)0.9814 (0.9690-0.9938)1.5195 (1.2087-1.9102)IV0.21 (−0.92 to 1.34)0.19 (−0.22 to 0.60)1.0021 (0.9908-1.0134)1.0019 (0.9978-1.0060)0.9998 (0.9879-1.0119)1.0754 (0.8450-1.3685)CI, confidence interval; IRR, incidence rate ratio; TNM, tumour–node–metastasis.Figure 3National trends in incidence pre- and post-RAPC implementation by TNM stage at diagnosis, stratified by deprivation quintile subgroups. RAPC, Rapid Access Prostate Clinic; TNM, tumour–node–metastasis. Table 3Relative trend in incidence rates (slope IRR) pre- and post-implementation by TNM stage at diagnosis, stratified by deprivation quintile subgroupsDeprivation quintileTNM 7 stagePercentage change in incidence rate per quarterSlope IRRChange in slopeImmediate level changePreimplementation periodImplementation periodPreimplementation periodImplementation period% (95% CI)% (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)IRR (95% CI)1I3.03 (2.24-3.83)0.18 (−0.11 to 0.46)1.0303 (1.0224-1.0383)1.0018 (0.9989-1.0046)0.9723 (0.9643-0.9802)0.9911 (0.8342-1.1773)II1.50 (0.52-2.48)−0.56 (−0.90 to −0.21)1.0150 (1.0052-1.0248)0.9944 (0.9910-0.9979)0.9798 (0.9698-0.9899)1.2711 (1.0276-1.5723)III1.02 (−0.25 to 2.30)1.00 (0.61-1.39)1.0102 (0.9975-1.0230)1.0100 (1.0061-1.0139)0.9998 (0.9867-1.0131)1.1899 (0.9304-1.5219)IV0.11 (−1.62 to 1.87)0.75 (0.19-1.32)1.0011 (0.9838-1.0187)1.0075 (1.0019-1.0132)1.0064 (0.9882-1.0250)1.2968 (0.8664-1.9409)2I3.82 (3.17-4.48)0.11 (−0.16 to 0.39)1.0382 (1.0317-1.0448)1.0011 (0.9984-1.0039)0.9643 (0.9576-0.9710)0.8904 (0.7769-1.0204)II3.11 (2.10-4.12)−0.67 (−1.05 to −0.28)1.0311 (1.0210-1.0412)0.9933 (0.9895-0.9972)0.9634 (0.9533-0.9736)0.8682 (0.7132-1.0568)III1.74 (0.62-2.87)0.41 (0.02-0.81)1.0174 (1.0062-1.0287)1.0041 (1.0002-1.0081)0.9870 (0.9755-0.9986)1.3484 (1.0208-1.7811)IV−0.43 (−1.63 to 0.79)0.29 (−0.27 to 0.86)0.9957 (0.9837-1.0079)1.0029 (0.9973-1.0086)1.0072 (0.9938-1.0208)1.3627 (0.9516-1.9512)3I3.82 (3.14-4.50)0.11 (−0.19 to 0.40)1.0382 (1.0314-1.0450)1.0011 (0.9981-1.0040)0.9643 (0.9573-0.9713)0.9376 (0.8099-1.0854)II1.87 (1.04-2.71)−0.78 (−1.14 to −0.42)1.0187 (1.0104-1.0271)0.9922 (0.9886-0.9958)0.9740 (0.9653-0.9827)1.2854 (1.0697-1.5446)III1.42 (0.68-2.17)0.52 (0.12-0.93)1.0142 (1.0068-1.0217)1.0052 (1.0012-1.0093)0.9911 (0.9829-0.9995)1.5495 (1.2297-1.9525)IV−1.10 (−2.55 to 0.36)0.57 (−0.06 to 1.20)0.9890 (0.9745-1.0036)1.0057 (0.9994-1.0120)1.0169 (1.0008-1.0333)1.3441 (0.9117-1.9814)4I3.93 (2.99-4.88)−0.03 (−0.33 to 0.28)1.0393 (1.0299-1.0488)0.9997 (0.9967-1.0028)0.9619 (0.9527-0.9712)0.9189 (0.7735-1.0916)II2.28 (1.46-3.11)−0.85 (−1.18 to −0.53)1.0228 (1.0146-1.0311)0.9915 (0.9882-0.9947)0.9693 (0.9609-0.9779)1.0362 (0.8603-1.2480)III2.89 (1.86-3.93)0.28 (−0.15 to 0.72)1.0289 (1.0186-1.0393)1.0028 (0.9985-1.0072)0.9746 (0.9640-0.9854)1.3030 (1.0108-1.6797)IV−0.61 (−2.00 to 0.80)0.35 (−0.29 to 1.00)0.9939 (0.9800-1.0080)1.0035 (0.9971-1.0100)1.0097 (0.9942-1.0255)1.2191 (0.8459-1.7569)5I3.89 (3.04-4.75)−0.23 (−0.50 to 0.04)1.0389 (1.0304-1.0475)0.9977 (0.9950-1.0004)0.9604 (0.9520-0.9688)0.9890 (0.8403-1.1641)II2.83 (1.79-3.89)−1.16 (−1.52 to −0.80)1.0283 (1.0179-1.0389)0.9884 (0.9848-0.9920)0.9612 (0.9508-0.9717)1.1728 (0.9560-1.4386)III2.00 (0.77-3.25)0.10 (−0.26 to 0.46)1.0200 (1.0077-1.0325)1.0010 (0.9974-1.0046)0.9814 (0.9690-0.9938)1.5195 (1.2087-1.9102)IV0.21 (−0.92 to 1.34)0.19 (−0.22 to 0.60)1.0021 (0.9908-1.0134)1.0019 (0.9978-1.0060)0.9998 (0.9879-1.0119)1.0754 (0.8450-1.3685)CI, confidence interval; IRR, incidence rate ratio; TNM, tumour–node–metastasis. Relative trend in incidence rates (slope IRR) pre- and post-implementation by TNM stage at diagnosis, stratified by deprivation quintile subgroups CI, confidence interval; IRR, incidence rate ratio; TNM, tumour–node–metastasis. National trends in incidence pre- and post-RAPC implementation by TNM stage at diagnosis, stratified by deprivation quintile subgroups. RAPC, Rapid Access Prostate Clinic; TNM, tumour–node–metastasis. For stage II disease, incidence increased across all quintiles during the preimplementation period, followed by consistent declines during implementation in every deprivation group, with the steepest reductions observed among more deprived quintiles. In contrast, stage III incidence continued to rise modestly during implementation across most quintiles. Stage IV incidence showed small positive trends following a largely stable preimplementation period, with no strong deprivation gradient evident. In analyses restricted to cancer centres (Supplementary Table S3 and Figure S2, available at https://doi.org/10.1016/j.esmorw.2026.100781), preimplementation increases were steeper across all stages and deprivation quintiles. Following implementation, stage I incidence stabilised in deprivation quintiles 1-4 and declined modestly in the most deprived quintile only. Stage II incidence flattened across quintiles and stage III and stage IV cancers continued to increase modestly, with no strong socioeconomic gradient evident. Stage-specific Wald tests assessing whether the RAPC-related change in incidence trends differed across deprivation quintiles showed no evidence of socioeconomic heterogeneity for any stage in analyses on the national level. In contrast, analyses restricted to cancer centres demonstrated significant heterogeneity across deprivation quintiles for stage I disease only [χ2 (4) = 18.45, P = 0.0010]. No significant differences in RAPC-related slope change were observed across deprivation quintiles for stages II-IV within cancer centres. Sensitivity analyses Findings were robust across all sensitivity analyses. Shifting the intervention date to Q1 2010 or Q1 2011, as well as exclusion of the implementation period (2009-2011) with Q1 2012 defined as the intervention point, produced results consistent with the primary analyses (Supplementary Tables S4-S9, available at https://doi.org/10.1016/j.esmorw.2026.100781). Exclusion of the COVID-19 period (2020-2022) did not materially alter estimates of postimplementation incidence trends (Supplementary Tables S10 and S11, available at https://doi.org/10.1016/j.esmorw.2026.100781). Similarly, inclusion of unstaged tumours as a separate category yielded findings comparable to the primary stage-specific analyses, indicating that improvements in staging completeness over time did not materially influence the observed trends (Supplementary Tables S12 and S13, available at https://doi.org/10.1016/j.esmorw.2026.100781). Adjustment for calendar quarter identified evidence of seasonal variation in quarterly diagnoses [Wald χ2 (3) = 14.47, P = 0.0023]. However, inclusion of calendar quarter had a negligible impact on the estimated RAPC-related level and slope changes (Supplementary Tables S14 and S15, available at https://doi.org/10.1016/j.esmorw.2026.100781). Discussion This study provides a national assessment of how diagnostic patterns changed following the introduction of RAPCs and whether RAPC-related changes differed by socioeconomic deprivation. Overall, the findings indicate that RAPC implementation coincided with substantial changes in incidence trends and diagnostic setting, alongside shifts in stage distribution. These changes occurred in the context of progressive centralisation of PCa diagnosis into designated cancer centres. However, these changes were not observed uniformly across socioeconomic strata, suggesting that changes observed following centralisation occurred within the context of pre-existing socioeconomic differences in access to early detection. The pronounced rise in PCa diagnoses in all deprivation quintiles before RAPC implementation is consistent with the widespread uptake of opportunistic PSA testing in Ireland in the absence of an organised screening programme.18 The subsequent attenuation of incidence trends following implementation aligns with national policy discouraging population-based screening, as reflected in National Cancer Forum recommendations,8 instead favouring symptom-based assessment and structured GP referral pathways.10 From this perspective, RAPC implementation occurred alongside a broader shift away from opportunistic testing, providing important context for postimplementation trends. Consequently, the changes in incidence trends observed in this study should be interpreted as being temporally associated with RAPC implementation rather than attributable solely to RAPCs themselves. Examining differences in incidence trends by deprivation quintile during the RAPC implementation period suggests that diagnostic entry may have played a greater role than downstream service centralisation in shaping observed patterns. The most marked postimplementation declines in incidence were observed among men living in more deprived areas, whereas incidence stabilised or continued to increase among less deprived groups. These divergent patterns are unlikely to reflect differences in access to RAPCs, which operate within publicly funded cancer centres, but instead suggest that changes in diagnostic activity were driven by socially patterned uptake of PSA testing and referral. This interpretation is supported by established evidence that PSA testing in routine health care settings is less common among men with lower income19 and lower educational attainment,20 implying that policy-driven reductions in opportunistic testing may have disproportionately reduced diagnostic entry among more deprived groups. Differences in private-sector utilisation across quintiles may also have contributed to these patterns. Less deprived men were more frequently diagnosed outside the public cancer-centre pathway, potentially reflecting greater access to private health care services. Stage-specific patterns further suggest that socioeconomic differences in diagnostic entry may have influenced the population-level patterns observed following centralisation. The postimplementation stabilisation of early-stage (stages I and II) incidence did not occur uniformly across deprivation strata, with declines concentrated among more deprived men and continued increases among less deprived groups. This pattern is consistent with international evidence showing that PSA-detected cancers are disproportionately concentrated among men of higher SES,21 and that higher SES is associated with lower risk of advanced disease primarily through greater engagement with early detection.22 Recent studies have similarly reported that men living in more socioeconomically deprived areas are less likely to undergo PSA screening23 and more likely to present with high-risk or metastatic PCa at diagnosis.24 However, in the present study, socioeconomic differences were more apparent for overall incidence and early-stage diagnoses than for late-stage presentation. These findings should also be interpreted in the context of PSA-driven overdiagnosis,25 described in Irish epidemiological trends of increasing incidence alongside relatively stable mortality.26 From this perspective, reductions in early-stage diagnoses among more deprived groups may partly reflect reduced detection of indolent, PSA-detected cancers, while continued increases among less deprived men may reflect ongoing overdiagnosis, given higher uptake of opportunistic PSA testing among more affluent men and the established link between PSA testing and overdiagnosis.19,25 The continued increase in stage I diagnoses observed within cancer centres also suggests that centralisation of diagnostic services does not necessarily eliminate the potential for overdiagnosis. In the absence of information on Gleason score or International Society of Urological Pathology Grade Group, it was not possible to determine whether the observed shifts in stage at diagnosis were accompanied by corresponding changes in tumour aggressiveness. Although RAPCs and associated national referral pathways were intended to improve the timeliness of diagnostic assessment within cancer centres, we found little evidence of a corresponding decrease in late-stage disease following their implementation. Patterns in stage III and IV disease were less pronounced, with only modest differences observed across socioeconomic groups, both nationally and within cancer centres. These findings suggest that socioeconomic differences were more apparent for early-stage diagnosis than for advanced disease presentation. Nevertheless, stage III and IV incidence continued to increase across the population throughout the implementation period, indicating that substantial challenges in achieving earlier diagnosis remain despite improvements in diagnostic organisation. While RAPCs may streamline diagnostic pathways once men enter specialist care,27 their population-level impact is constrained by socioeconomic inequalities in diagnostic entry. These findings are directly relevant to ongoing European efforts to redesign PCa early detection. The European Union-funded PRostate cancer Awareness and Initiative for Screening in the European Union (PRAISE-U) initiative28 is evaluating approaches to organised PCa screening across several Member States, including Ireland, with the aim of improving equitable access to diagnostic entry.29 However, evidence indicates that population-based screening programmes may not fully address underlying socioeconomic inequalities in uptake and can introduce challenges related to overdiagnosis. Alternative approaches, including risk-stratified screening for high-risk groups, have been proposed as a means of improving equity in early detection while minimising potential harms.30 Implications in practice The introduction of national GP referral guidelines, alongside the centralisation of diagnostics through RAPCs, coincided with a stabilisation of early-stage detection following a period of opportunistic, PSA-driven expansion in diagnosis. However, access to RAPCs remains dependent on upstream processes, particularly PSA testing. Consequently, changes in patterns of opportunistic PSA testing may have disproportionately reduced early-stage diagnosis among more deprived men. The implications of this reduction are uncertain: it may reflect reduced overdiagnosis of indolent cancers or reduced detection of clinically important early-stage disease. If the former, less deprived groups, who are more likely to undergo PSA testing, may be more exposed to overdiagnosis despite guidance discouraging opportunistic testing. These findings indicate that diagnostic centralisation alone is insufficient to achieve equitable and optimal PCa detection. Strengths and limitations A key strength of this study is the use of data from the NCRI, a high-quality, population-based registry with complete national coverage and statutory reporting of incident cancer diagnoses.31 Analyses were not age-standardised within the 50-70−year age range. However, the age distribution of cases was highly similar before and after RAPC implementation, making confounding by age unlikely. Information on ethnicity and race was not available and therefore could not be incorporated into the analyses. Following RAPC introduction, 3.8% of PCa records had unknown stage at diagnosis and 8.1% had missing deprivation information. However, sensitivity analyses including unstaged cases produced findings comparable to the primary analyses and deprivation was evenly distributed among cases with known SES, suggesting limited risk of bias. Changes in TNM staging over time may affect stage comparability. However, PCa cases diagnosed between 2000 and 2013 were retrospectively recoded to TNM7 using available clinical T, N, and M15 data to maximise consistency in longitudinal analyses. Information on Gleason score and International Society of Urological Pathology Grade Group was not available in the registry data. Consequently, we were unable to assess whether observed changes in stage distribution were accompanied by changes in tumour aggressiveness. As an uncontrolled interrupted time-series study, this analysis cannot attribute observed changes in incidence trends and stage distribution solely to RAPC implementation. RAPCs were introduced during a period of broader health care reform that included centralisation of cancer services, implementation of national referral guidelines, and evolving PSA testing practices, all of which may have contributed to the observed associations. Finally, deprivation-specific population denominators were not available for the full study period. Deprivation-stratified analyses therefore compared temporal patterns in incident case counts using a population offset rather than estimating deprivation-specific incidence rates. Findings should therefore be interpreted as evidence of differences in incidence trends between deprivation groups over time, rather than precise estimates of deprivation-specific incidence rates. Conclusion The introduction of RAPCs was associated with changes in PCa incidence and stage distribution among Irish men aged 50-70 years and achieved substantial centralisation of diagnostic care. However, these effects were not uniform across socioeconomic groups. Reductions in early-stage (stages I and II) incidence were greatest among more deprived men, while advanced-stage incidence continued to increase across the population, indicating that streamlined access to specialist diagnostics alone does not eliminate late presentation. These findings suggest that the population-level impact of diagnostic centralisation is likely shaped by upstream socioeconomic gradients in PSA testing and referral. Addressing inequities in pathway entry will be essential if improvements in diagnostic organisation are to translate into equitable gains in PCa outcomes.