Changes in PSA-Based Early Detection of Prostate Cancer Over a 12-Year Period: Findings from the German KABOT Study

1. Introduction

Prostate cancer is the most frequently diagnosed malignancy among men in Germany and the United States, with approximately 65,820 and 268,500 newly diagnosed cases per year, respectively [1,2]. It represents the second leading cause of cancer-related death in men in both countries, accounting for 15,379 deaths annually in Germany and 34,500 in the United States [1,2]. Incidence rates peak in men in their eighth decade of life, whereas prostate cancer mortality is highest among men aged 80 years and older [2,3]. Globally, prostate cancer remains the most common cancer and the fifth leading cause of cancer-related death among men, with an estimated 1.4 million new cases and 375,000 deaths reported worldwide in 2020 [1,2,3].

Marked disparities in prostate cancer incidence and outcomes have been consistently observed. African American men exhibit the highest incidence rates worldwide [4]. A positive family history constitutes another well-established, non-modifiable risk factor [4]. Several potentially modifiable factors have also been associated with prostate cancer incidence, including high alcohol consumption, dietary calcium and milk protein intake, and hormonal influences [4,5,6]. In contrast, higher consumption of tomatoes, increased physical activity, intake of phytoestrogens such as those found in soy products, higher coffee consumption, and greater sexual activity have been discussed as protective factors [7,8,9,10,11]. Notably, current smoking status has repeatedly been associated with a lower reported incidence of prostate cancer [12]. This paradoxical finding has been attributed to lower participation in prostate cancer early detection among smokers rather than a true protective effect. Large population-based data have demonstrated that active smokers are approximately one third less likely to undergo prostate-specific antigen testing compared with non-smokers, even after multivariable adjustment [13]. Importantly, despite lower detection rates, smokers exhibit a significantly increased risk of prostate cancer-specific mortality [14].

Prostate-specific antigen testing was introduced in 1994 as an adjunct to digital rectal examination for the early detection of prostate cancer [15]. Potential harms of PSA-based early detection include complications related to subsequent prostate biopsy as well as the risks and long-term consequences of overtreatment [16]. Although the effectiveness of PSA-based early detection remains controversial, robust evidence supports a reduction in prostate cancer-specific mortality in defined age groups. The European Randomized Study of Screening for Prostate Cancer demonstrated a 27% reduction in cancer-specific mortality after a median follow-up of 21 years, despite substantial non-compliance and contamination in both study arms [17]. Conversely, population-level data from the United States indicate that the discontinuation of routine PSA-based screening has been accompanied by an unfavorable stage migration. While incidence rates among men aged 45 to 74 years remained stable between 2004 and 2010, a significant increase in metastatic prostate cancer was observed from 2010 onward following changes in screening recommendations [18].

In 2012, the United States Preventive Services Task Force recommended against routine PSA-based screening [19]. This position was revised in 2018, endorsing shared decision-making for men aged 55 to 69 years [20]. Subsequent studies from the United States consistently reported a sustained decline in PSA testing following these changes, accompanied by reductions in prostate biopsy rates, prostate cancer incidence, and the use of definitive local therapies [21,22,23]. These findings highlight the sensitivity of early detection practices to guideline recommendations and health policy decisions.

In Germany, early detection of prostate cancer is performed by both urologists and general practitioners and is offered from the age of 45 years. Men with statutory health insurance are entitled to an annual digital rectal examination; however, this examination does not constitute a screening test and has limited sensitivity for early-stage disease [24]. Measurement of PSA levels is not routinely reimbursed and is performed only upon patient request after counseling regarding potential benefits and harms [25]. Consequently, PSA-based early detection in Germany largely depends on individual awareness, physician counseling, and patient-driven decision-making.

The effectiveness of any screening strategy depends not only on its diagnostic performance but also on its implementation within the target population. Beyond age, sex, and ethnicity, several patient-related factors, including educational attainment, income, insurance status, and marital status, have been shown to substantially influence screening behavior [21,22,26,27,28,29,30,31,32,33,34]. Recent epidemiological studies have further demonstrated that PSA testing uptake is closely linked to broader health behaviors and indicators of preventive care engagement, underscoring the complex interplay between sociodemographic factors and early detection practices [35].

In parallel to the ongoing controversy surrounding PSA-based early detection, European urological initiatives have increasingly shifted the focus from test-based screening towards structured, risk-adapted early-detection pathways [36]. These concepts aim to preserve potential mortality benefits while reducing harms through improved pre-biopsy risk stratification, integration of magnetic resonance imaging, and consistent use of active surveillance for low-risk disease [36,37,38]. Within this framework, digital rectal examination is no longer considered a stand-alone screening tool, but rather an adjunctive clinical assessment with limited discriminatory value when used in isolation [37]. Similar to population-based breast cancer screening, the emerging consensus emphasizes that the value of early detection depends less on the screening test itself than on the quality, structure, and governance of the diagnostic pathway in which it is embedded [36].

Identifying population subgroups with limited awareness or reduced utilization of PSA-based early detection is therefore essential for targeted counseling and informed decision-making. The present study evaluates awareness and utilization of PSA-based early detection of prostate cancer over a 12-year period, enabling an assessment of temporal changes in the context of evolving evidence and international guideline recommendations. Of particular interest is whether the decline in PSA testing observed in the United States following the 2012 guideline change is similarly reflected in a German healthcare setting.

In addition to quantifying utilization patterns, this study assesses knowledge and awareness of PSA-based early detection among men considered eligible for testing. Given the lack of comparable longitudinal studies spanning more than a decade, this pilot study aims to provide a data-driven foundation for future large-scale investigations into early detection behavior and targeted counseling strategies.

2. Materials and Methods

2.1. Implementation of the Study

In 2009, an interdisciplinary working group consisting of general practitioners and urologists conducted a cross-sectional survey study to assess attitudes, knowledge, and utilization patterns related to prostate cancer screening and early detection based on prostate-specific antigen testing. A structured questionnaire comprising 20 items was developed, refined, and validated through 12 structured individual interviews to ensure clarity, comprehensibility, and content validity.

Study Phase 1 (SP1) was conducted in 2009. A total of 55 general practitioner practices were randomly selected from a database provided by the regional Association of Statutory Health Insurance Physicians in Berlin and Brandenburg. Each participating practice received 50 questionnaires for distribution to male patients aged 35 years or older who were considered capable of autonomously completing the survey.

Study Phase 2 (SP2) was initiated in October 2021 to evaluate temporal changes over a 12-year interval. This follow-up survey was conducted under the study label “Knowledge and Belief Over Time (KABOT)”. The questionnaire content remained largely identical to SP1, with the addition of two thematic blocks addressing contemporary aspects of prostate cancer early detection. The moderate response rate observed in SP1 of 38.2% raised concerns regarding sample size and representativeness. To address these limitations, the number of contacted general practitioner practices was increased to 150 in SP2.

Practices in SP2 were selected using an updated random sampling approach and were complemented by practices affiliated with the KV RegioMed teaching network, identified using predefined geographic search criteria (general practitioner practices located in or around Cottbus or Bernau within a 5 km radius) [39,40]. A reminder letter was distributed in January 2022. All questionnaires returned by March 31, 2022, were included in the final analysis. Data collection was performed by physicians accredited within the statutory health insurance system, while exclusively private practices were excluded. All surveys were completed anonymously in full compliance with applicable data protection regulations.

Ethical approval was obtained from the Ethics Committee of the Brandenburg State Medical Association (approval number 2021-2126-BO-ff) on August 10, 2021. The study was registered in the German Registry of Clinical Studies (DRKS registration number 00027862) [41].

2.2. Structure of the Questionnaires

Questionnaire development followed the Consensus-Based Checklist for Reporting of Survey Studies (CROSS), with full adherence documented in Table A1 and Table A2 [42]. The patient questionnaire used in SP1 included 11 items addressing sociodemographic characteristics, family history, and vaccination status, as well as nine additional items assessing knowledge, perceptions, and personal utilization of prostate cancer early detection (Appendix A1).

In SP2, one additional item was included to assess participants’ perceptions regarding the potential impact of early detection on cancer-specific mortality (Appendix A2). Pilot testing was performed in both study phases with 12 participants each, resulting in minor adjustments to optimize survey length, internal consistency, clarity, and completeness.

Response rates were calculated at the practice level and refer to practices that actively participated by returning completed patient questionnaires.

2.3. Research Questions and Statistical Methods

The primary objective of this study was to identify factors independently associated with prostate cancer early detection awareness and utilization. For clarity and focus, analyses were restricted to early detection of prostate cancer.

Age was recorded using predefined categorical response options, including a category for participants younger than 35 years, followed by 5-year increments from 35–39 years up to a category of 75 years and older. Smoking status was recorded as non-smoker, former smoker, or current smoker and was dichotomized for regression analyses into “non-smoker or former-smoker” as the reference category versus “current smoker”. Educational attainment was assessed using four categories reflecting the German educational system: basic secondary school qualification, secondary school certificate, higher school certificate, and university degree. For analysis, these were dichotomized into “basic secondary school qualification or secondary school certificate” versus “higher school certificate or university degree”.

Family history of cancer was recorded as “none”, “one case”, or “several cases” and subsequently dichotomized into “no family history of cancer” versus “family history of cancer”. Paternity was categorized as “no children” versus “children”. Employment status options included “jobseeker”, “employed”, “self-employed”, and “pensioner”, and were dichotomized into “jobseeker” versus “employed, self-employed, or pensioner”. Insurance status was recorded as statutory or private health insurance. Study phase (SP1 versus SP2) was included as an independent variable in all analyses.

Two predefined binary endpoints were analyzed: awareness of prostate-specific antigen-based early detection and actual utilization of prostate-specific antigen-based early detection. Univariable logistic regression models were first applied to assess the individual association of each independent variable with the endpoints. Subsequently, multivariable logistic regression models were constructed to identify independent predictors while accounting for potential confounding. Variables were included in the multivariable models based on clinical relevance and statistical significance in univariable analyses, with all dichotomized variables entered using the lower-risk or absence category as the reference group. In case of an incompletely answered questionnaire, available answers were nevertheless considered and appropriately reflected in the analysis.

Comparisons of baseline characteristics between SP1 and SP2 were performed using chi-squared tests for categorical variables and independent-samples t-tests for continuous variables. Effect estimates are reported as odds ratios with corresponding 95% confidence intervals (95% CI). Statistical significance was defined as a two-sided p-value of 0.05 or less. All statistical analyses were conducted using SPSS software version 29.0 (IBM Corp., Armonk, NY, USA).

3. Results

The composition of the study population across both study phases is illustrated in the flow chart shown in Figure 1. The analytic cohort comprised 755 participants in Study Phase 1 (SP1) and 135 participants in Study Phase 2 (SP2) (total n=890) (Table 1).

3.1. Participant Characteristics

Baseline characteristics stratified by study phase are summarized in Table 1. Compared with SP1, SP2 included a higher proportion of non-smokers (63.0% vs. 48.5%, p<0.001) and a higher proportion of participants with a university degree (38.5% vs. 30.5%, p=0.002). Employment status distributions differed between SP1 and SP2 (p<0.001), with a markedly higher share of self-employed participants in SP2 (45.2% vs. 4.9%). Paternity was less frequent in SP2 (80.0% vs. 85.4%, p=0.019). In contrast, differences by study phase were not statistically significant for age distribution (p=0.068), family history of cancer (p=0.102), or insurance status (p=0.344).

3.2. Awareness of PSA-Based Early Detection

Univariable and multivariable logistic regression results for the endpoint “Awareness of PSA-based early detection” are presented in Table 2. In univariable analyses, awareness increased with age (OR 1.063, 95% CI 1.005–1.124; p=0.032) and was higher among participants with higher educational attainment (OR 2.021, 95% CI 1.496–2.730; p<0.001) and among participants with children (OR 2.081, 95% CI 1.414–3.064; p<0.001). Current smoking was significantly associated with lower awareness of PSA-based early detection (OR 0.553, 95% CI 0.397–0.770; p<0.001). Employment status was also associated with awareness (OR 1.829, 95% CI 1.108–3.018; p=0.018). Study phase was not associated with awareness in univariable analysis (OR 1.028, 95% CI 0.699–1.513; p=0.888).

In the multivariable model, higher educational attainment (OR 1.709, 95% CI 1.239–2.357; p=0.001), smoking status (OR 0.688, 95% CI 0.479–0.989; p=0.043), and paternity (OR 1.939, 95% CI 1.249–3.012; p=0.003) remained independently associated with awareness, whereas age (OR 0.990, 95% CI 0.926–1.058; p=0.766) and employment status (OR 1.528, 95% CI 0.888–2.631; p=0.126) were no longer significant after adjustment. Family history of cancer and insurance status were not independently associated with awareness.

3.3. Utilization of PSA-Based Early Detection

Regression results for the endpoint “Already underwent PSA-based early detection” are summarized in Table 3. In univariable analyses, utilization was higher in SP2 than SP1 (OR 1.476, 95% CI 1.022–2.131; p=0.038) and increased with age (OR 1.417, 95% CI 1.327–1.513; p<0.001). Higher educational attainment (OR 1.873, 95% CI 1.422–2.469; p<0.001), paternity (OR 2.646, 95% CI 1.718–4.076; p<0.001), and employment status (OR 2.473, 95% CI 1.402–4.362; p=0.002) were positively associated with utilization, whereas smoking was inversely associated (OR 0.378, 95% CI 0.264–0.541; p<0.001).

In the multivariable model, increasing age (OR 1.391, 95% CI 1.291–1.498; p<0.001) and higher educational attainment (OR 1.632, 95% CI 1.190–2.239; p=0.002) remained positively associated with utilization, while current smoking remained independently associated with reduced utilization of PSA-based early detection (OR 0.608, 95% CI 0.405–0.912; p=0.016). Study phase was no longer significant after adjustment (OR 1.365, 95% CI 0.885–2.105; p=0.160). Paternity and employment status also lost statistical significance in the multivariable model.

3.4. Subgroup Analysis in Men Aged 45–69 Years

Subgroup analyses restricted to men aged 45–69 years are shown in Table 4 and Table 5. For awareness of PSA-based early detection (Table 4), higher educational attainment was associated with awareness in univariable analysis (OR 1.875, 95% CI 1.258–2.794; p=0.002) and remained significant after adjustment (OR 1.717, 95% CI 1.113–2.647; p=0.014). Current smoking was independently associated with lower awareness of PSA-based early detection in both univariable (OR 0.525, 95% CI 0.343–0.804; p=0.003) and multivariable analyses (OR 0.595, 95% CI 0.375–0.945; p=0.028). Study phase was not associated with awareness in this subgroup.

For utilization of PSA-based early detection in men aged 45–69 years (Table 5), age was strongly associated with utilization in univariable analysis (OR 1.528, 95% CI 1.338–1.745; p<0.001) and remained significant in the adjusted model (OR 1.520, 95% CI 1.315–1.757; p<0.001). Higher educational attainment was also independently associated (OR 1.590, 95% CI 1.077–2.347; p=0.020). Current smoking remained associated with a lower likelihood of utilization in the adjusted model, reaching borderline statistical significance (OR 0.627, 95% CI 0.393–1.001; p=0.051). Smoking showed a borderline association after adjustment (OR 0.627, 95% CI 0.393–1.001; p=0.051). Study phase, employment status, family history of cancer, and insurance status were not independently associated with utilization in this subgroup.

Across all multivariable models (Table 2, Table 3, Table 4 and Table 5), study phase, employment status, family history of cancer, and insurance status were not independently associated with either endpoint.

4. Discussion

This longitudinal survey study provides several novel insights into prostate-specific antigen based early detection behavior in a non-organized screening setting over a 12-year period. First, despite substantial changes in international screening recommendations and accumulating evidence on the benefits and harms of PSA-based early detection, utilization patterns in Germany remained remarkably stable over time. Second, stable and pronounced social gradients persisted, with educational attainment and smoking behavior emerging as the most consistent determinants of both awareness and utilization of PSA-based early detection. Third, increasing age was strongly associated with utilization but not with awareness, indicating a widening gap between knowledge and action across the life course. Together, these findings add a longitudinal, population-based perspective to the international literature and highlight persistent structural determinants of early detection behavior.

The absence of a measurable study phase effect contrasts sharply with data from the United States, where the 2012 recommendation of the U.S. Preventive Services Task Force against routine PSA screening was followed by a sustained decline in PSA testing across all age groups [21,22,43,44,45,46,47,48,49]. Several studies have further demonstrated downstream consequences, including reductions in prostate biopsies, prostate cancer incidence, and definitive local therapy [23], accompanied by a subsequent increase in advanced and metastatic disease [18,47]. In particular, Desai et al. reported a significant rise in metastatic prostate cancer among men aged 45–74 years after 2011, whereas incidence rates had remained stable between 2004 and 2010 [18]. In contrast, our data indicate that these international developments did not translate into reduced PSA-based early detection use in Germany. This discrepancy likely reflects fundamental differences in health care organization, reimbursement structures, and the absence of centralized screening policies in Germany, where PSA testing remains largely patient-driven and dependent on individual counseling.

Our findings of stable awareness and utilization of PSA-based early detection over a 12-year period should be interpreted in the context of rapidly evolving European early-detection concepts led by urologic stakeholders [36,37,38]. Recent European initiatives emphasize that PSA testing alone is insufficient to generate population-level value and must be embedded within structured, risk-adapted algorithms incorporating downstream triage tools such as magnetic resonance imaging and defined surveillance strategies [36,38]. This paradigm mirrors the experience from population-based mammography screening, where quality assurance, standardized pathways, and centralized governance were essential to balance benefits and harms despite persistent debate on overdiagnosis [36]. In contrast, the opportunistic nature of PSA testing in Germany may partly explain why evolving evidence and recommendations have not translated into measurable changes in early-detection behavior within the present study.

Educational attainment emerged as the most robust determinant of PSA-based early detection behavior in this study. Higher education was independently associated with both awareness and utilization, in the overall cohort as well as in men aged 45–69 years. These findings are consistent with prior studies demonstrating higher PSA testing rates among men with higher educational levels [28]. At the same time, conflicting evidence exists. Pickles et al. reported higher PSA testing rates among men with lower educational attainment and limited health literacy, attributing this pattern to insufficient understanding of overdiagnosis and overtreatment [32]. Taken together, these findings underscore that educational level does not uniformly translate into more appropriate screening decisions, but rather modulates how information is processed and acted upon. From a public health perspective, our results suggest that men with lower educational attainment represent a key target group for structured, comprehensible counseling that facilitates informed and self-determined decision-making.

Smoking status constituted another consistent determinant of PSA-based early detection behavior. In both the overall cohort and the age-restricted subgroup, smokers showed significantly lower awareness and utilization of PSA-based early detection, with smoking remaining an independent predictor after multivariable adjustment. This finding aligns closely with population-based data from Golijanin et al., who demonstrated that active smoking was associated with approximately one third lower odds of PSA testing among more than 50,000 screening-eligible men [27]. These behavioral patterns also provide a plausible explanation for epidemiological observations linking smoking to a lower reported incidence of prostate cancer during the PSA screening era [12], despite a clearly increased risk of prostate cancer-specific mortality among smokers [14]. Together, these data indicate that reduced engagement in preventive health behavior, rather than biological protection, underlies this association and identifies smokers as a particularly vulnerable group for targeted counseling interventions.

Paternity was associated with higher awareness of PSA-based early detection but did not translate into higher utilization after adjustment. This dissociation may reflect greater general health awareness among men with children, without a corresponding increase in decisive action. Similarly, age showed a strong and independent association with utilization but not with awareness. This pattern suggests that while knowledge about PSA-based early detection remains relatively stable across age groups, actual uptake increases with advancing age. This finding is particularly relevant given robust evidence from the ERSPC demonstrating a reduction in prostate cancer-specific mortality in men aged 50–65 years [17]. It underscores the need to ensure that counseling is not deferred until older age, but rather reaches men at a time when the potential benefit of early detection is greatest.

A range of additional determinants of PSA-based early detection behavior has been described in the literature, including marital status, income, insurance status, and family history of cancer. An overview of selected findings is provided in Table 6.

In contrast to some prior reports [21,22,26,27,28,29,30,31,32,33,34], we did not observe an independent association of family history of cancer or employment status with awareness or utilization of PSA-based early detection. Insurance status was likewise not associated with PSA testing behavior, despite the fact that PSA testing is not routinely reimbursed by statutory health insurance in Germany but is commonly covered by private insurance. These findings suggest that financial coverage alone may not be sufficient to overcome informational and behavioral barriers to early detection.

The strengths of this study include its extended observation period of 12 years, the assessment of both awareness and actual utilization of PSA-based early detection, and the use of consistent methodology across study phases. To our knowledge, no comparable longitudinal survey addressing these aspects over such a time span has been published to date. Several limitations must nevertheless be acknowledged. The most important limitation is the relatively small sample size in Study Phase 2, which coincided with the coronavirus pandemic and may have affected participation. The smaller sample size in Study Phase 2 limits statistical power for detecting modest temporal effects and may increase the risk of type II error in subgroup analyses. Differences in participant characteristics between SP1 and SP2, particularly with respect to smoking behavior, educational attainment, and employment status, raise concerns regarding cohort comparability. In addition, recruitment was limited to patients visiting general practitioner practices, introducing potential selection bias toward individuals with higher health care utilization. The limited number of participating physicians further restricts generalizability, and it remains unclear to what extent participating physicians actively supported early detection and counseling. Finally, the questionnaire did not capture the depth or quality of participants’ knowledge regarding potential harms of PSA-based early detection, precluding a more nuanced assessment of informed decision-making.

In summary, this descriptive and hypothesis-generating pilot study demonstrates persistent social gradients in PSA-based early detection behavior over a 12-year period, despite evolving international evidence and recommendations. Educational attainment and smoking status emerged as the most consistent determinants of both awareness and utilization, while no temporal decline in PSA testing was observed in the German non-organized screening setting. These findings underscore the need for targeted, evidence-based counseling strategies that address structural and behavioral barriers to informed participation in early detection. Future implementation research should focus on linking patient knowledge, screening behavior, and downstream diagnostic quality within structured urologic pathways, in order to better evaluate how early detection strategies translate into clinically meaningful and equitable outcomes.

5. Conclusions

In this 12-year longitudinal survey, utilization of PSA-based early detection of prostate cancer remained stable in a German, non-organized screening setting despite substantial changes in international evidence and screening recommendations. Educational attainment and smoking status consistently emerged as the strongest determinants of both awareness and utilization, whereas increasing age primarily influenced utilization but not awareness.

These findings highlight a persistent gap between knowledge and action and underscore the relevance of social and behavioral determinants for early detection practices. Men with lower educational attainment and active smokers represent key populations for targeted, evidence-based counseling aimed at supporting informed and self-determined decision-making.

As a descriptive and hypothesis-generating pilot study, these results provide a robust foundation for future research integrating patient knowledge, screening behavior, and clinical outcomes, particularly in light of recently demonstrated mortality benefits of PSA-based screening in defined age groups.

Appendix A1. Questionnaire SP1 (Original Questionnaire in German, Translation for Publication)

Appendix A2. Questionnaire SP2 (Original Questionnaire in German, Translation for Publication)